Friday, March 22, 2013

List of Ways Reserves Leave the Banking System

See this post for a simplified list of where reserves can go (ignoring the distinction between inside and outside the banking system).

First of all, we need to define what is meant by reserves in the banking system. In the US this commonly includes reserves in the private banks' central bank (Federal Reserve Bank or just "Fed") reserve accounts (these are the only accounts on which the Fed pays "interest on reserves" (IOR)). It also includes "vault cash" which is physical currency/cash (paper bills and coins) stored at the private banks. It does not include the Federal government's Fed reserve account (the Treasury General Account (TGA)). It also does not include anything the Fed itself holds as an asset: reserves are never a Fed asset; they are always a Fed liability. All Fed reserve accounts are electronic. The Fed is an independent hybrid public/private institution, and thus not strictly part of the government. Of course the Treasury Department (or just "Treasury") is part of the federal government. We also need to consider other government agency Fed accounts (government sponsored enterprises, or GSEs).

Here, then, is the list of ways in which reserves leave the private banking system:
  1. When entities pay taxes/tariffs/fees/fines (and those payments are transferred to the TGA or GSE Fed accounts)
  2. When Treasury or GSEs auction bonds (and the proceeds are transferred to the TGA or GSE Fed accounts)
  3. When private non-banks withdraw paper bills and coins (physical cash or currency) from their private bank deposits
  4. When private banks repay Fed reserve loans or overdrafts (principal and interest)
  5. When foreign central banks or institutions (e.g. the IMF) receive funds in their Fed reserve accounts
  6. When the Fed/Treasury/GSEs sell assets (typically Treas. bonds, but also could include foreign currency, TARP assets, etc.) to private entities, for example during Fed Open Market Sales (OMSs)
First of all notice that "loaning out reserves to private non-banks" is not on the list. That's an incorrect way to think of reserves. Only chartered banks, Treasury, and certain GSEs, foreign central banks, and institutions (IMF, World Bank, etc.) can hold Fed reserve deposit accounts. Individuals, non-bank businesses, and most organizations cannot hold Fed reserve accounts. Reserves can be loaned by individual banks to other banks, and they can be transferred between banks to back/clear purchases or payments between private entities including when the purchasing entities are the banks themselves (e.g. to pay bank employee salaries, or to buy office supplies for the bank), but these events do not cause reserves to leave the banking system as a whole.

I've not made a distinction between excess reserves (ER) and required reserves (RR) in the above, although you can take the list to apply to ERs since RRs are of course required (by regulation), and thus are only absent (below their required levels) for brief periods of time. There are other ways in which, in aggregate (i.e. taking all banks as a whole), excess reserves can be converted to required reserves (and thus in that sense "leave the banking system"), but those are not covered by the above list since it makes no distinction between reserve types. Of course there are other ways reserves can leave, come into, or be converted from excess to required status by individual banks (as opposed to the banks in aggregate).

Of the six entries on the list, the first two and the interest component of the fourth* are typically reversed when the Federal government spends money. The third is reversed when currency is re-deposited in private banks by private entities. The principal component of the fourth is reversed by the Fed loaning out reserves (note that reserve loans are typically made on the inter-bank market, but the Fed stands ready as a lender of last resort, thus the bulk of Fed reserve loans are typically repaid in short order when replacement reserves are obtained from other entities). The fifth is reversed when funds exit these foreign central banks and institutions (note that foreign owned private banks are part of the banking system). The sixth is reversed by Fed Open Market Purchases (OMPs). Together OMSs and OMPs constitute Fed Open Market Operations (OMOs). So in terms of reserves permanently leaving the banking system, this only happens when the Federal government starts accumulating money in the TGA (i.e. running a surplus), when currency is permanently kept, destroyed or lost by the private sector, when Fed reserve loans are repaid, or when OMSs are not eventually reversed by OMPs. Government surpluses have historically been rare and comparatively little currency is kept, destroyed or lost by the private sector. Net "permanent" Fed reserve loans are really only made to the aggregate banking system to support reserve requirements (typically ~10% of bank demand/checking deposit liabilities). This leaves OMSs as the main way in which reserves leave the banking system most of the time. Conversely, OMPs are the main way reserves are injected into the banking system most of the time. OMPs are how "Quantitative Easing" (QE) is accomplished. Treasury deficit spending does NOT inject reserves into the banking system. Instead it injects net financial assets (i.e. Treasury bonds) into private hands, and at the same time takes the proceeds from the bond auctions and spends them back into the private sector. Thus, in a sense (though not literally), the Treasury obtains and spends private bank created money (inside money) when it deficit spends. This is not literally true, since the Treasury spends from it's Fed deposit (the TGA), but this deposit was funded largely through the private sector: either by taxes or bond auctions. Non-banks purchasing Treasury debt or paying taxes use inside money to do so. In the case of private non-banks, funding the TGA is the mirror image of spending from the TGA: funding involves the elimination of a bank deposit and the simultaneous transfer of outside money from the banks (which may have to borrow from the Fed for this purpose) into the TGA. Spending involves the transfer of outside money from the TGA to the banks, and the simultaneous creation of bank deposits for the non-bank payees. Of course what counts as money can be debated. In our system, bank deposits have a very high degree of moneyness.

Assuming that Treasury immediately spends its Fed deposit (i.e. the TGA is immediately emptied: not a terrible assumption with deficit spending), then the sum of non-bank held commercial bank deposits (& cash) is equal to the sum of private bank loans and the Treasury debt held by the Fed & banks. Why is this? (henceforth in this discussion I'll use "bank deposits" to mean "bank deposits & cash" for simplicity, since cash is mostly just withdrawn deposits). For starters, since "loans create deposits" we'd expect to be able to express the bank deposits of non-banks in terms of loans (loans to Treasury being called "Treasury debt" here). But why doesn't this statement include all Treasury debt? To see why consider the following example: assume we start off with everyone's balance sheets clear (zeros for assets and liabilities), and then the Fed & banks acquire $X in Treasuries. This results in $X in bank deposits when Treasury spends. Now these deposits can be traded by the non-banks for the Treasuries and back again.... in any amount up to $X. Thus bank deposits can vary between $0 and $X under this trade. Now assume that bank loans in the amount of $Y are made to non-banks. Now non-bank deposits can vary between $Y and $(X+Y) depending on the amount of Treasuries traded with the Fed & banks (the maximum still being $X). At this point we've accounted for, in terms of debt, all existing non-bank bank deposits. If the non-banks now use any part of these deposits to purchase Treasuries from Treasury, those deposits will be returned to them (in aggregate) once Treasury spends and thus the aggregate bank deposits will not change. Since this process can be repeated an indefinite number of times (assuming Treasury continues to auction bonds and spends all the proceeds each time), the total amount of T-bonds acquired by the non-banks in this manner can total $Z, where Z can be any positive value, without changing the aggregate non-bank bank deposits. If the Fed & banks purchase some of these new Treasuries from the non-banks, the non-bank bank deposits will increase accordingly. Thus the total amount of bank deposits held by the non-banks can vary between $Y and $(X+Y+Z) depending on how many Treasuries are traded with the Fed & banks. Only to the extent that the Fed & banks hold a part of the $(X+Z) in existing Treasuries, are non-bank bank deposits elevated above $Y, and thus the only case in which these bank deposits total the full $(X+Y+Z) is when the Fed & banks purchase all existing Treasury debt.

Note that I'm simplifying a bit here by glossing over the process by which Treasury Tax and Loan (TT&L) accounts are created as an intermediate step before transferring funds to the TGA -- this transfer being necessary before the funds can be spent by Treasury.

These JKH and Ramanan comments at monetaryrealism.com provides further insight

*Note that almost all the interest paid to the Fed for its reserve loans is remitted by the Fed to Treasury, and thus will almost certainly be spent again into the private sector.

123 comments:

  1. Very interesting and helpful info Tom! I was particularly intersted in your explnation of the difference between deficit spending and increasing reserves in the banking system.

    There's a lot of confusion about what govt deficit spending is-and isn't. If you have anything further to elaborate on this or know sources that do it'd be appreciated.

    I recall how in the 80s everyone thought that Reagan's huge military Keynesian infused deficits-really they were the products of both the military buildup and the huge tax cuts-would lead to "galloping inflation."d

    Of course, nothing of the kind happened. I think that's a related matter that gets confused a lot: what causes and doesn't cause inflation. Whatever you think of huge deficits-my trouble with Reagan's deficits are allocational; they were fueled by the tax cuts and military buildup- they don't cause inflation.

    ReplyDelete
    Replies
    1. Mike, that bit on deficit spending is pure Cullen Roche! He likes to put it as "gov deficit spending is just redistributing inside money (from Peter to Paul) and issuing an NFA in the process (to Peter)."

      It only occurred to me later to qualify that a little bit. I make an attempt at that in the 2nd to last paragraph above starting with "Only to the extent that..."

      Actually I thought a little bit more about that today. In that paragraph you'll notice that I qualify my conclusion by assuming the TGA is empty. I'm trying to come up with a more general way to state the idea there. In our system there's only two entities capable of actually creating dollars "ex-nihilo" (out of thin air) (I'm ignoring the fact that Treasury mints physical cash... since the Fed is the institution that trades it to/from the banks). Those entities are the private banks, through creating loans, and the Fed through purchasing (or perhaps lending too... not sure about this). Thus these dollars created must equal the sum total of dollars "out there" (both in and outside of the banking system). In particular this includes the TGA + private bank deposits + cash + money banks hold. I'm not sure how to handle "money banks hold" since banks are unique in the way they purchase items (see my Example #5). Joe in Accounting (on pragcap) has pointed out that sometimes banks make use of "correspondent banks" to handle their own purchases, but I concluded that those could safely be ignored, since it's not required, and it doesn't change the big picture.

      Anyway, as you can see I haven't fully fleshed this idea out yet, but you'll notice that what's missing here is the government creating dollars ex-nihilo. The government ends up USING dollars either created by the banks or the Fed (mostly the banks), but it doesn't create them. It can, however, create T-bonds which are not dollars (they fall a little further down the scale on Cullen's "moneyness" spectrum).

      Anyway, if you make a few simplifying assumptions (TGA = $0, and ignore banks' "own money"), I think my 2nd to last paragraph is basically correct in the post: That is: only to the extent that the Fed holds treasuries, can you trace dollars in the private sector to gov debt. All the rest is bank created. So if the Fed didn't hold any treasuries, all our $ would be bank created, regardless of how big the gov debt was.

      Delete
  2. Tom, nice blog - useful and clear.

    I disagree with some of the points in the above post, however...

    1) "The Fed is an independent hybrid public/private institution, and thus not strictly part of the government"

    The Fed is actually part of the government:

    http://www.usa.gov/Agencies/Federal/Independent.shtml

    Here's what the Fed says about itself:

    "The Federal Reserve System is the central bank of the United
    States. It was founded by Congress in 1913 to provide the
    nation with a safer, more flexible, and more stable monetary
    and financial system."

    “The Federal Reserve System is considered to be an independent central bank because its decisions do not have to be ratified by the President or anyone else in the executive branch of government. The System is, however, subject to oversight by the U.S. Congress. The Federal Reserve must work within the framework of the overall objectives of economic and financial policy established by the government; therefore, the description of the System as “independent within the government” is more accurate”.

    “Congress designed the structure of the Federal Reserve System to give it a broad perspective on the economy and on economic activity in all parts of the nation. It is a federal system, composed of a central, governmental agency—the Board of Governors—in Washington, D.C., and twelve regional Federal Reserve Banks”.

    “The Board of Governors of the Federal Reserve System is a federal government agency. The Board is composed of seven members, who are appointed by the President of the United States and confirmed by the U.S. Senate”.

    “Congress chartered the Federal Reserve Banks for a public purpose. The Reserve Banks are the operating arms of the central banking system, and they combine both public and private elements in their makeup and organization. As part of the Federal Reserve System, the Banks are subject to oversight by Congress”.

    http://www.federalreserve.gov/pf/pdf/pf_complete.pdf

    ReplyDelete
    Replies
    1. Thanks phil. I won't argue w/ your quotes above. I think what I'm getting at is that there is a designed in separation between the Fed and the (rest of) government. As it stands now the Fed cannot directly purchase bonds from Treasury, nor can it forgive the principal for the T-bonds it does acquire (it does remit the interest payments). The econviz website has an example of a hypothetical means of government self financing where the Fed is allowed to purchase Treasury bonds directly. That's on their "macro" page under "Government Spends (Without Borrowing)" operation. They stress that's only a hypothetical operation and doesn't correspond to anything that actually happens.

      Delete
    2. Phil, there's also this (very short write up):

      http://www.federalreserve.gov/faqs/about_14986.htm

      So the member banks own stock in the Fed, and they must as a requirement of membership. It is unusual stock though. I accept this "independent within government" description. So you can strike my "hybrid" comment if you wish, but I think it helps get the independence of the institution across. I think that's certainly the way the Fed has behaved over the past few decades: it's hard to argue that it isn't there to serve the interests of the private banks given this behavior.

      Also keep in mind that I consider the whole point of this post to be a little bit arbitrary: what's in and outside the banking system has more to do with arbitrary definitions than anything else. The really important points I'm trying to get across to people is that reserves don't flow into the general public. For that to happen the general public would have to be allowed to have reserve accounts. The other point is in my final paragraphs concerning the origin of "money out there in the economy." The VAST majority of that money can be said to have originated from private debt. Say there's about $65 T of dollars in existence in the economy, and the TGA is always running on empty (and again I'm ignoring TT&L) due to deficit spending. Then public debt doesn't matter. What matters is the $ of public debt owned by the Fed. They own about 10% (of the total public debt), or about $1.6T or so. Thus only about 2.5% of "money out there in the economy" can be said to have originated from PUBLIC debt! The other 97.5% is money that originated from private debt! So when a private individual buys a T-bond (say through Treasury direct), then the money that gets recycled back into the economy through the process of deficit spending was, on average, 97.5% privately created.

      Delete
    3. "it's hard to argue that it isn't there to serve the interests of the private banks given this behavior".

      That doesn't contradict it being a part of the government. Government policy is currently to let the Fed try and "macro-manage" the economy by playing around with interest rates and quantitative easing, and to keep the banking sector afloat at all costs.

      "reserves don't flow into the general public"

      My understanding is that reserves are basically federal reserve notes in electronic form. Banks withdraw their deposits from the Fed as cash (notes), and the public withdraws cash (notes) from their banks... so the public gets its hands on "reserves" in the form of cash.

      another comment follows...

      Delete
    4. You write: "Banks withdraw their deposits from the Fed as cash (notes), and the public withdraws cash (notes) from their banks... so the public gets its hands on "reserves" in the form of cash."

      My understanding is that vault cash held at banks are considered reserves. Once the public withdraw portions of their deposits as cash, this cash ceases to be reserves. So yes, that's true. In that sense reserves do flow into the public through that channel. That's #3 on my list. So perhaps that statement is too broad. But people aren't withdrawing any more cash than is convenient to do. My argument is that this escape route for reserves is minor and does not represent a major pathway. That pathway is reversed when cash is re-deposited back at the bank, and thus re-acquires its reserve status.

      The point is that all those excess reserves (in electronic form) due to QE do not flow out to the public. For that too happen the public would have to have reserve accounts. In other words, other than the amount withdrawn as cash, banks don't "lend out" those excess reserves to the public. That's a common misconception people have and is what really inspired this post.

      Delete
    5. "Thus only about 2.5% of "money out there in the economy" can be said to have originated from PUBLIC debt!"

      The non-bank public can buy government bonds if it wants to, but logically the means to purchase those bonds must be provided by either Fed lending or purchases, or by previous Treasury spending. In fact, the only way the non-govt can own reserves, with which it can purchase bonds, is if the government has previously spent in deficit. Otherwise the non-govt simply has to borrow from the Fed.

      So essentially what happens is that banks either borrow from the Fed or use the proceeds from previous government spending to buy govt bonds.

      Government spending then results in a credit to bank reserve accounts and to non-bank deposit accounts.

      The non-bank agents receiving this spending then usually spend their deposits. At some point some other non-bank agent then decides to save, perhaps by buying a government bond (from a bank).

      So, government spending initially results in credits to bank reserve accounts and to non-bank deposit accounts. Those that receive this money then decide what to do with it. Either they spend it, or decide to save it (perhaps by buying a govt bond).

      At any point in time, however, those bank deposits in existence which can be said to be 'backed by government spending' will generally be equal, I think, to the quantity of bonds or reserves held by banks.

      Delete
    6. "My understanding is that vault cash held at banks are considered reserves. Once the public withdraw portions of their deposits as cash, this cash ceases to be reserves".

      You're right, cash in bank vaults counts as bank reserves.

      If banks have more paper/coin cash reserves in their vaults than they need, they can deposit them at their local Fed bank (just as we can deposit cash at our local bank). And they can also withdraw their deposits from the Fed in the form of cash (just as we can withdraw our deposits from our banks as cash).

      In general the non-bank paper/coin cash requirement is pretty constant, so excess bank reserves at the Fed don't stimulate the public to take out more money from the ATM.

      However, if the Fed were to impose a negative interest rate on excess bank reserve balances (as some have suggested they should), then banks would probably just withdraw more of their Fed deposits as cash and just keep it in their vaults (if they were allowed to of course).

      Delete
    7. "At any point in time, however, those bank deposits in existence which can be said to be 'backed by government spending' will generally be equal, I think, to the quantity of bonds or reserves held by banks".

      Actually, thinking about it, this is incorrect. At the same time I'm pretty certain that "only to the extent that the Fed holds Treasury bonds as assets can you say that private bank deposits in the private sector are backed by "loans to the government" is also incorrect. It would be useful to get some actually statistics on this, as it seems like 'murky water' to me.

      Delete
    8. phil, I had a response written but then erased it because I can't recall exactly where I saw that $65T in $ in the private sector figure. It occurred to me to check the following chart (linked to at the bottom of Example 4):

      http://www.zerohedge.com/news/2012-12-26/record-2-trillion-deposits-over-loans-feds-indirect-market-propping-pathway-exposed

      The 1st chart on that page (for some reason the page doesn't come up for me right now). The ~$9T in deposits they show there jives with this chart:

      http://upload.wikimedia.org/wikipedia/en/5/58/MB%2C_M1_and_M2_aggregates_from_1981_to_2012.png

      So was the $65T a measure of M3 or M4... I'm not sure!

      The point is I'm suddenly a little uncertain about my 2.5% figure, but regardless of the actual percentages, this is my reasoning, based on the idea that all dollar denominated money comes from debt:

      Banks have a charter to issue dollar denominated deposits which act as money for most of us. Of course there's cash too, but more and more deposits are being used directly. The debt which created that money is clearly the bank loan. Every time you swipe your credit card this kind of money/debt pair is created. Buying a house or car with borrowed funds is the same.

      Now the only other entity which can create money out of thin air like that is the Fed. The Treasury can create bonds out of thin air (as do businesses), and they are certainly low risk, but the bonds are not the same as money. The Fed creates perfectly liquid money through a debt mechanism very similar to how private banks do:

      The Fed either creates it by loaning it as reserves, or purchasing assets with it (typically Treasuries). In either case there's a debt instrument they can put under their assets column when they do that, perfectly offsetting the freshly created reserves.

      The government does not have a mechanism for doing precisely this (ignoring coins). Only the banks and the Fed can create money this way. Ignoring physical cash for the moment, the difference is that the Fed creates money to be used by only the banks and the Treasury (reserves), and the banks create money for the rest of us (deposits).

      Delete
    9. For simplicity lets create a world with no required reserves, a single commercial bank, and no taxes or government spending, and no physical cash. The entire private sector can run just fine without any reserves at all. There's really no purpose for them: The bank just credits and debits its customers' deposits to realize all payments between customers. Of course all the deposits were originally loaned into existence to another set of customers. Now fold in a second bank, and suddenly SOME payments require the Fed to TEMPORARILY loan reserves, but its balance sheet can be clear again by the end of each day (See my examples #1 and #1.1). The money supply expands and contracts according the the $ amount of outstanding loans the private banks have. Now let's add to this world government taxing and spending. Still no problem. Again the Fed must jump in to facilitate the process, but assuming the gov spends every dollar it taxes, then the consolidated balance sheets look exactly the same before taxing as they do after spending. Now lets add in gov deficit spending. Again the sector consolidated balance sheets (think econviz) look exactly the same before and after the process except that private non-banks now have T-bond assets from the gov... but the amount of dollar denominated MONEY still matches precisely the principal amount of the outstanding private bank loans (assuming the banking sector essentially operates for free and doesn't collect retained earnings). All through these steps the Fed acts as a facilitator and its balance sheet is clear before and after each operation. Only during intermediate steps is its BS populated. That changes when the Fed buys T-bonds from the public. Now the deposits held by the public grows, but this growth is exactly offset by T-bonds sold to the Fed. Suddenly the dollar denominated deposits used by the public can be said to be a mix of Fed created money and bank created money: The amount that's Fed created is exactly equal to the value of the bonds it holds. The amount that's bank created is the same as it was before the Fed bought T-bonds. The amount created by the Fed can be said to have originated in government debt... but only that debt held by the Fed. The rest of the deposits in the private system were created by private debt (loans) held by the banks. It's a given, of course, that the total amount of T-bonds out there must exceed the amount held by the Fed, but other than that the total $ amount of T-bonds has no influence on the amount of actual $ denominated money in the system. Again every time I'm stopping the system to examine it here are times at which the TGA = 0 again (as it originally did all the time before we added gov spending). I think that's the clearest time to sample.

      Delete
    10. Tom,

      "Banks have a charter to issue dollar denominated deposits which act as money for most of us. Of course there's cash too, but more and more deposits are being used directly."

      Even if the whole world stopped using cash, this wouldn't change the "hierarchical" nature of the monetary system. The quantity of cash in circulation is not really relevant.

      "The Treasury can create bonds out of thin air (as do businesses), and they are certainly low risk, but the bonds are not the same as money".

      Currency is a debt of the government, and bonds are a debt of the government which pays interest. Federal Reserve notes and Treasury bonds are both US government securities. There is no involuntary default risk on US government bonds. They are widely referred to as "cash- equivalents", "money-good", or even just "cash". They are among the most liquid assets in the world. Saying they are not money is rather pointless. The reality is that Treasuries are for all intents and purposes savings accounts at the Fed.

      Also, if you agree with the idea that there is a sliding scale of "moneyness", then Treasuries must be a form of money.

      "The government does not have a mechanism for doing precisely this (ignoring coins). Only the banks and the Fed can create money this way"

      The Fed is part of the government. Its power to issue currency comes from Congress. The currency it issues is a liabiltiy of the government.

      The government has the power to create money at will if it wants to, but it limits what the Treasury can do in this regard. The Treasury creates notes and coins, but only issues coins. The Treasury does have a mechanism for creating money - coins. There is no reason to ignore this.

      Government spending is logically money creation. The money the government spends ('outside money) is a liability of the government. This means the government issues/creates liabilities (money) when it spends, and extinguishes liabilties (money) when it taxes.

      "Ignoring physical cash for the moment"

      Why do you ignore whatever doesn't fit with your argument? Physical cash exists, there is no need to ignore it.

      Delete
    11. I agree that Treas bonds have a high degree of moneyness, but still are not precisely money. We could do an experiment purchasing items from brick and mortar and internet businesses... with credit cards, personal checks, cash, and Treasury bonds. Which do you think would be accepted as money the least?

      OK, fine I was using "government" as shorthand for non-Fed government. The currency and reserves the Fed issues show up on the Fed's balance sheet as liabilities don't they? Not the non-Fed government's balance sheet.

      Ignoring coins: fine, don't ignore the coins. I still think they are relatively insignificant in the grand scheme of things.

      "Government spending is logically..."

      I think it's more clear to separate the Fed, which keeps its own balance sheet, from the rest of government (the non-Fed gov... and assume "Treasury" keeps it BS). Treasury thus spends from its assets (reserves) not by creating liabilities. It cannot spend if it has no reserve assets. The Fed, however, can buy assets by creating reserves as liabilities from thin air. When the Fed sells assets, the reserve liabilities it created out of thin air are extinguished. When the non-Fed gov taxes, private bank deposits are debited by the same amount that Treasury reserve assets and increased. The private bank is on the hook for coming up w/ the reserves to transfer to Treasury. If it doesn't have the reserves it can borrow from other banks or the Fed to do this. If it borrows from the Fed, the Fed actually CREATES reserve liabilities (on its BS) to do this, while taking the loan to the bank (for the reserves) as an asset on its BS. In either case, no reserves are "extinguished" when the non-Fed gov taxes. If the bank has the reserves or borrows them from another bank, those reserves are TRANSFERRED to Treasury! If the bank borrows from the Fed, the Fed CREATES them and then transfers them to Treasury.

      "physical cash": OK, fine, don't ignore it.

      Delete
    12. phil, take a look here:

      http://econviz.org/macroeconomic-balance-sheet-visualizer/

      and run the operations "Government Taxes" and "Government Spends"

      With the former you'll see the following (and ONLY the following):

      Household deposit-assets decrease by $30
      Banks reserve-assets decrease by $30
      Treasury's reserve-assets increase by $30

      No reserve-liabilities are "extinguished" in this process. In this case the "banks" had enough reserves to back the tax payment, so no borrowing from the Fed was required, and the "Central Bank" balance sheet remains unchanged.

      The exact mirror image of this operation is the latter ("Government Spends"). Run it and you'll see the following:

      Household deposit-assets increase by $30
      Banks reserve-assets increase by $30
      Treasury's reserve-assets decrease by $30

      No government government liabilities are "created." In particular, the central bank's balance sheet is again completely unchanged. Treasury reserve-assets decrease.

      Delete
    13. I shouldn't have said "and ONLY the following," because, of course, the banks' deposit-liabilities change the same amount and direction as the banks' reserve-assets, and the households' equity changes the same amount and direction as the households' deposit-assets. That goes for both the "Tax" and "Spend" operations.

      Delete
    14. "We could do an experiment purchasing items from brick and mortar and internet businesses..."

      Can you use a $100 note to buy a chocolate bar? In most cases, no. The Fed used to issue $10,000 notes - could you use these in a shop? No. That doesn't mean they're not money. Bonds aren't used as money in small denomination transactions, but they are equivalent to money within certain markets, and can be converted into bank deposits or 'cash' at a moment's notice. Treasury bills in particular have such a high degree of liquidity that the distinction between them and 'cash' in this regard is largely irrelevant, hence they are usually referred to as cash-equivalents. To large financial institutions and foreign governments T-bills are basically just 'cash'.

      http://www.investopedia.com/terms/c/cashequivalents.asp

      According to JKH, short-term US Treasuries are actually more liquid than reserves:

      "short term debt and reserves at the Fed are not perfect substitutes. Short term debt is more liquid. Banks can’t “sell” reserves to non-banks. The market for bills is much broader than the market for bank reserves."

      http://www.interfluidity.com/v2/3694.html

      If we define "moneyness" as liquidity, this would mean that T-bills actually have a higher degree of "moneyness" than reserves.

      Whether you call Treasuries money or not, of course depends on how you define ‘money’. If you take a very narrow and conventional view, then they are not. If you take a slightly broader view, then they are.

      Delete
    15. phil, thanks for the links/quotes. I did, however, already agreed with you that there's a "moneyness" scale and that Treasuries do rank high on that scale. I agree with most of what you say here. For a bank, sure I can see that short term debt is "more liquid." Of course they can't directly clear payments with them or transfer them to the TGA, but you're correct that they can easily sell them for reserves before doing that. In regards to JKH's quote "Bank's can't *sell* reserves to non-banks" ... that's true, but banks can transfer reserves to the banks that non-banks bank at to pay for things that non-banks are selling! (See my example #5) ... which of course increases the non-bank's deposits at their (reserve receiving) banks too.

      I still think that at MOST businesses that we private non-bank individuals, businesses and organizations purchase items from, you'll find that Treasury bonds would be the least accepted as money (from my list). I would guess in fact that either checks or credit cards would be the MOST widely accepted as money. You already outlined why physical cash isn't always accepted by businesses in large denomination bills. I don't even think the IRS or the states accept physical cash for tax payments, do they?
      And yes I know not all businesses accept personal checks or credit cards. I forgot to mention bank cards (debit cards). You can use those at an ARCO station or COSTCO but not a VISA or Mastercard card!

      Personally I use bank cards, credit cards, and cash mostly... and rarely a personal check. None of them are PERFECT money, but they all have higher "moneyness" than T-bonds. If you want to argue that T-bonds have a higher moneyness than reserves... fine. I still think there's an argument against that, however bank deposits and physical cash are really king, aren't they? When was the last time you made a purchase with a T-bond? Ever had a direct deposit from your employer made in T-bonds? I don't think if I asked a 100 friends, family and acquaintances if they'd ever purchased anything or accepted payment from anything in T-bonds that any of them would say "Yes." So lets start with you: Have you ever paid or been paid in T-bonds? Perhaps I'll start with my morning coffee purchase today... I'll ask if I can pay with T-bonds. What kind of a response do you think I'll get? When I make purchases for my work, I sure don't use T-bonds either. Everything I and everybody I know use on a daily basis is either physical cash (obtained mostly from bank deposits: i.e. ATM machines), or a bank deposit related mechanism: credit card, debit card, or personal check. Occasionally I use a money order or get one for payment.

      You helped me make my point about reserves actually, because the majority of entities don't use them as money (physical cash outside the bank, remember is NOT reserves), nor do they accept T-bonds as money. However, regarding *accepting* reserves as money, the public probably does w/o knowing it: every time a bank pays for something or when you get a tax refund check from the gov, or when the gov or bank pays it's employees. To the public those payments look like payments into their bank deposits.

      I think Cullen's "moneyness" scale has a reasonable ordering, although I know some items are debatable. I'd like to see it presented more in a table format... with each kind of money on a different row and green check boxes in columns representing where that money is recognized as such.

      Delete
    16. Economist JP Koning (from whom Cullen got the idea of "moneyness"), argues that Treasury bonds are money. Both he and economist Stephen Williamson agree with JKH that T-bills are more liquid than reserve balances (I assume this is a broadly held understanding), which by JP Koning's definition means that they have more "moneyness" than reserves.
      It's important to note however that people define "moneyness" in slightly different ways. JP Koning defines it simply as liquidity:
      http://jpkoning.blogspot.co.uk/2013/03/ranking-moneyness.html
      whilst economist Perry Mehling describes it as the 'quality' of a given type of credit within a gradated hierarchical structure, in which "What looks like money at one level of the system looks like credit from the standpoint of the level above":
      ieor.columbia.edu/pdf-files/Mehrling_P_FESeminar_Sp12-02.pdf
      (Mehrling's paper is really worth a read by the way. He also discusses inside and outside money).
      "The currency and reserves the Fed issues show up on the Fed's balance sheet as liabilities don't they? Not the non-Fed government's balance sheet".
      Fed liabilities are counted as assets on the Treasury's balance sheet, but they are also liabilities (referred to as obligations) of the US government, so this is slightly misleading. When the Treasury has a positive balance in its account at the Fed, the US government is holding its own liability, but this is shown as an asset on the Treasury's BS and a liability on the Fed's BS. Basically the two are different sides of a consolidated government balance sheet, in which the asset and liability net to zero. The balance in the Treasury's account can be thought of as showing how many government liabilities have been returned to the government, i.e. extinguished, (through taxes, etc). (I say extinguished because when you return a liability to the issuer the liability is extinguished).

      Now this seems a bit too conceptual for some people, but it’s perfectly logical, and factual. Fiat money is fundamentally some sort of liability, obligation, or other such “promise” of the government or State. Different countries account for this in slightly different ways, but this basic fact is always the same. In the US this fact is made clear within the legal definition of terms, but is obscured slightly by institutional arrangements. For example, in the US the Treasury is required by the government to have a positive balance in its account before it spends, and to account for this balance as an “asset”. This doesn’t change the fact that the balance represents a government obligation by definition.

      As far as I’m concerned the statements “the Treasury needs to get money before it can spend” and “the government creates money when it spends” are completely compatible. They are different ways of describing the same thing.

      By the way Brad Delong (economist and former assistant deputy secretary of the Treasury) agrees with the description that government spending "creates money" and taxes "destroying" money: http://delong.typepad.com/sdj/2013/03/bill-black-is-justifiably-irate-monday-hoisted-from-comments-weblogging.html


      “fine, don't ignore the coins. I still think they are relatively insignificant in the grand scheme of things”

      You’re right that the role of coins in the economy today is minor. Originally, however, coin was considered the “ultimate” form of money, with everything else just a form of credit. As it says in the Constitution: “The Congress shall have Power To… coin Money, regulate the Value thereof”.

      Congress could issue coins of any denomination if it wanted to, and the Treasury is currently legally permitted to mint platinum coins of any face value. As such, though coins are generally only used as small change, they remain a form of money which shouldn’t be ignored.

      Delete
    17. Sorry I posted the above comment before I'd seen your comment!

      Delete
    18. "I still think that at MOST businesses that we private non-bank individuals, businesses and organizations purchase items from, you'll find that Treasury bonds would be the least accepted as money (from my list). I would guess in fact that either checks or credit cards would be the MOST widely accepted as money"

      You're using a particular definition of money or moneyness. As I said people don't use T-bonds in small transactions, but to large financial institutions and to wholesale money markets T-bills are basically "money". To everyone else they are "cash equivalents" which can be converted at the drop of a hat into cash/deposits, as they are extremely liquid.

      Usually, checks and credit cards are not considered to be "money", though checks do occasionally circulate as currency substitutes. You don't give someone a credit card in payment, rather the credit card instructs your bank to pay on your behalf (as does a check). But again this is a matter of definitions in which you have both narrow conventional views and broader views.

      Money is a complex phenomenon and there is a lot of disagreement over the terminology surrounding it. What you call "money" or how you define "moneyness" depends on your theoretical framework and what you are trying to explain, so different views can potentially be compatible if they are taken in context.

      Similarly, any "rankings" of money or "moneyness" will depend on your definitions. Cullen's scale is based on his own particular view, which is not necessarily shared by economists who have written on the subject. For example JP Koning doesn't provide a definitive "ranking" of moneyness; as he says: "Moneyness, after all, is subjective... While market prices are surely the best way to build moneyness rankings, the problem is that markets rarely provide the prices to facilitate the analysis... There is no independent moneyness market in which an asset's liquidity services can be sliced away from all the other services, thereby allowing those liquidity services to be bought, sold, and priced".

      http://jpkoning.blogspot.co.uk/2013/03/ranking-moneyness.html

      And as I said other economists define moneyness differently.

      During the gold standard era, most of the currency in circulation was banknotes, and banks also created credit money on their books as they do today. So did banknotes or accounting entries have more "moneyness" than gold coins or gold bullion? Cullen would say that they did. Again, it depends on how you define moneyness.

      During the gold bullion standard gold didn't circulate at all within the economy, but the currency was still pegged to gold and gold was used in international transactions between national central banks. According to your definition, gold in this case had almost zero moneyness! Yet it was at the very center, and the foundation of, the entire monetary system.

      Delete
    19. phil, OK lets take your position and consolidate the Fed and Treasury balance sheets. This changes nothing about the econviz operations I referred you to. All you need to do is change the name from "Treasury" to "Consolidated Fed/Treasury" and everything else is the same. It's still the case that "Government Spends" results in "Consolidated" assets going down with no new liabilities created. And "Government Taxes" results in "Consolidated" assets increasing with no liabilities extinguished. This is exactly what happens in the "Federal Government Sector (aggregate)" balance sheet above both the Fed and Treasury on econviz. I agree w/ econviz here.

      I expect that your response will be that to truly consolidate/aggregate the balance sheets you must do some cancellations: reserves on the assets side must be cancelled with reserves on the liabilities side. Yes if you do that, what you say is correct.

      It comes down to what we view as what's "obscuring" what's going on. I'll grant you that viewing the Fed/non-Fed as a unified consolidated whole and aggregating the balance sheets in the way I'm supposing you would want results in your view. However, I think viewing it in that way obscures the very different nature of the Fed and non-Fed sides of things and their designed in separation,... and it obscures the reason why the Fed wasn't just created as a desk in the Treasury Dept. I think its more useful to view them as separate entities and you don't. That's where we disagree.

      Delete
    20. "Usually, checks and credit cards are not considered to be "money", though checks do occasionally circulate as currency substitutes. You don't give someone a credit card in payment, rather the credit card instructs your bank to pay on your behalf (as does a check). But again this is a matter of definitions in which you have both narrow conventional views and broader views."

      Of course! You don't think I meant that people circulate the physical checks or physical credit cards as a form of money do you?

      And I agree about the subjective nature. I think Cullen's moneyness scale is reasonable and you probably disagree.

      Gold bullion days: Yes, I suppose you are correct. It would be viewed as lower on my moneyness scale. It would be higher back when it actually circulated amongst the bulk of the population.

      Delete
    21. "I'll grant you that viewing the Fed/non-Fed as a unified consolidated whole and aggregating the balance sheets in the way I'm supposing you would want results in your view."

      You can consolidate the Fed and Treasury into one general government balance sheet, but you don't have to do that for the statement "government spending creates money" to make sense. If "outside money", or fiat money is a government obligation/liability (which is what it is - or at least the "outside money" which the government receives and spends is), then when the Treasury receives money (through taxes, etc), government obligations are simply returned to the government. This means that those government obligations are extinguished.

      Look at it this way:

      1. Write "IOU $10" on a piece of paper, then give that piece of paper to someone.

      2. Take your IOU back from that person.

      Your IOU has now been extinguished. The piece of paper still exists, but the IOU has been cancelled. The "money" is the IOU promise, not the piece of paper on which that promise is written.

      You're right that the Fed and Treasury are seperate entities, and that the Treasury has to have a positive balance in its account before it is allowed to spend. This is a government-imposed constraint on Treasury operations which doesn't change the above logic.

      A way of describing this is to say that the government has imposed upon itself the requirement that it only create as much money through Treasury spending as it has destroyed through taxes and bond sales.

      Alternatively you could say that the Treasury gets money and then spends it, given that the Treasury is a government department and not "the government".

      Both are different ways of describing the same process. To the Treasury the credits in its account are money which it gets and spends. To the government they are obligations which are destroyed and created. But they are only 'money' in the first place because they are government obligations.

      An important question is whether this constraint on Treasury operations places a limit on the government's ability to spend. The answer is that it doesn't (you seem to agree with this given that you're into MR). If for some reason it did, then it wouldn't exist for long if the government needed or wanted to spend. It's self-imposed.

      Delete
    22. Of course if you define Treasury bonds as a form of 'money' - which follows if you see money as a "scale of moneyness", even if you argue that bonds have less "moneyness" than deposits/cash/reserves - then the Treasury does also create 'money' when it issues bonds.

      Delete
    23. Tom,

      perhaps we should agree to disagree on the above and move on to the other points, if you're still interested in discussing them?

      Regarding your comment re the econviz "government taxes and government spends" demo:

      If the banks have reserves with which to pay taxes, then where did they come from?

      For the banks to own reserves which aren't simply borrowed from the Fed, either the govt has to have previously deficit spent, or the Fed must have purchased assets from the private sector.

      In either case, the government, either through Fed lending or purchases, or through previous deficit spending, has to provide the funds which are then used to pay taxes or to buy Treasury bonds.

      Delete
    24. phil,

      "perhaps we should agree to disagree on the above and move on to the other points, if you're still interested in discussing them?"

      Yes, lets do that.

      "For the banks to own reserves which aren't simply borrowed from the Fed, either the govt has to have previously deficit spent, or the Fed must have purchased assets from the private sector."

      "...borrowed from the Fed.": Check. I agree.
      "...Fed must have purchased...": Check. I agree.
      "...deficit spent...": I don't agree w/ that one.

      Again I'm defining deficit spending (as econviz does in it's "Government Spends (Consoliated)"), from the bond auction all the way through the proceeds (every $) being spent by Treasury (I'm tired of writing non-Fed gov, so I'm using "Treasury" for that concept from now on). Look at the econviz balance sheets. Reserve levels do not change for anybody. If you define "deficit spending" in another way (with different start and/or stop points), then perhaps it's different. I prefer the econviz definition. The consequences of that definition are plain to see on econviz.

      If the real point of your comment is to say "reserves must have come from the Fed" I completely agree. That's the ONLY place they could possibly come from. And I agree that they are always a Fed liability. Every reserve dollar out there is a Fed liability.

      Again, I think this just gets back to our view of the Fed vs Treasury. You see them as the same thing with one big consolidated balance sheet, and consolidated in such a way that reserves in the asset column must cancel reserves in the liability column, and I'll grant you that there's justification for seeing it that way, and if you do see it that way (most) of what you write here makes perfect sense! I prefer to see them as separate entities which maintain separate balance sheets. Viewed that way you cannot escape the consequences. I don't think we've really "moved on" here... the issue is the same. It all goes back to that one issue.

      Do you agree that my view is consistent w/ the econviz view? If you do, are you saying the econviz view is incorrect? If so, what in particular is wrong with it? How would you change it to correct it? That's the part I'm not getting. I think if you grant me that the Fed and Treasury are separate, then nothing I've written here is inconsistent with that econviz tool and it's easy enough to run it and see the consequences. In particular, my view leads to the conclusion that ONLY the Fed can create reserves. The Treasury and private banks can only USE reserves. The one exception being coins perhaps... yes I know Treas mints them and issues them as opposed to paper money which it mints and gives to the Fed to issue. I figured this peculiarity must have some historical basis as you stated in another comment... So I'll grant you that coins (not paper money) are a chink in the "armor" of my argument, but I still maintain they're an insignificant "chink" ... unless of course we do that $1T coin thing, which nobody seems terribly interested in doing (yet).

      So in summary, the special case of coins is not enough of a factor for me to abandon my view. When we start churning out $1T coins on a regular basis to fund Treasury, I'll be happy to change my view. Or if the Fed is fully stripped of its (semi) independence and it buys its stock back from the private banks, and it becomes just a desk at Treasury I will change my view. I think my view is perfectly consistent w/ the econviz view, and I think there are significant advantages to seeing it that way. So we're back to where we started I think. We just need to agree to disagree on that point and move on (which is the same point as before really!).

      Delete
    25. Hi Tom, sorry for the delay.

      I had a look at the econviz ‘macroeconomic balance sheet visualizer’. On step #6 (government issues debt) it says that “government bond issuance comes after spending and simply gives the private sector a longer duration asset in exchange for money (i.e. it’s an asset swap)”.

      What it means is this:

      1. Initial setup:

      Tsy: A: 40 fed deposits, L: 210 bonds, NE: 170
      Fed: A: 170 bonds, L: 40 tsy deposits, 80 reserves, 40 notes, E: 10
      Banks: A: 80 reserves, 40 notes, L; 80 deposits, E: 40
      Non-banks: A: 80 deposits, 40 bonds, L: 0, E: 120

      2. Treasury spends 40:

      Tsy: A: 0, L: 210 bonds, NE: 210
      Fed: A: 170 bonds, L: 120 reserves, 40 notes, E: 10
      Banks: A: 120 reserves, 40 notes, L; 120 deposits, E: 40
      Non-banks: A: 120 deposits, 40 bonds, E: 160

      (Private sector net financial wealth has increased by 40. Deposits and reserves have increased by 40).

      3. Treasury sells 40 bonds:

      Tsy: A: 40 fed deposits, L: 250 bonds, NE: 210
      Fed: A: 170 bonds, L: 40 tsy deposits, 80 reserves, 40 notes, E: 10
      Banks: A: 80 reserves, 40 notes, 40 bonds, L; 120 deposits, E: 40
      Non-banks: A: 120 deposits, 40 bonds, E: 160

      (Private sector net financial wealth is unchanged).

      4. Households buy 40 bonds:

      Tsy: A: 40 fed deposits, L: 250 bonds, NE: 210
      Fed: A: 170 bonds, L: 40 tsy deposits, 80 reserves, 40 notes, E: 10
      Banks: A: 80 reserves, 40 notes, L; 80 deposits, E: 40
      Non-banks: A: 80 deposits, 80 bonds, E: 160

      (Private sector net financial wealth is unchanged).

      Government spending (2) provides income and net financial assets to the private sector, increasing private sector net financial wealth. Households spend this income, and some households decide to purchase government bonds (4). This purchase is simply an asset swap which leaves private sector net financial wealth unchanged.

      Some people get confused by this and think that bond issuance increases private sector net financial assets, whilst government spending simply redistributes assets. But this is incorrect; it’s actually the government spending itself which increases net financial assets. This is always the case no matter how you look at it. Whether you have bond issuance before government spending or after makes no difference.

      Note that households (or non-banks) are not required to buy the bonds. They choose to buy the quantities they want given the quantities available. If they don’t want to buy them, then the bonds simply remain within the banking system. For example:

      4. Households buy 20 bonds:

      Tsy: A: 40 fed deposits, L: 250 bonds, NE: 210
      Fed: A: 170 bonds, L: 40 tsy deposits, 80 reserves, 40 notes, E: 10
      Banks: A: 80 reserves, 40 notes, 20 bonds, L; 100 deposits, E: 40
      Non-banks: A: 100 deposits, 60 bonds, E: 160

      (Private sector net financial wealth is unchanged).

      Delete
    26. phil, no argument with any of that except:

      On step #6 (government issues debt) it says that “government bond issuance comes after spending and simply gives the private sector a longer duration asset in exchange for money (i.e. it’s an asset swap)”.

      I'm not seeing that on econviz. The word "after" doesn't appear in the description for the operation on the pull down list entitled "Government Issues Debt," which happens to be the 4th item on the pull down list. Nor does it appear for "Government Spends" or "Government Spends (Consolidated)." I'm not sure what you mean by "step #6" but I assume you're talking about the ordering on this pull down list of operations. The 6th item does mention the word "after." That operation is entitled "Government Spends (Without Borrowing)." I don't see a sentence precisely like the one you've quoted, but I do see this:

      "That is, the consolidated government (treasury plus central bank) can always spend if desired without taxing or issuing bonds either beforehand or afterwards."

      Which is similar to what you quoted. However, note that it also says this:

      "Governments currently do not use this spending approach, but they could if desired."

      In other words "Government Spends (Without Borrowing)" is merely a hypothetical operation. It doesn't actually work that way.

      Otherwise, Treasury has to have funds in the TGA BEFORE it can spend. Perhaps, hypothetically, they could legally NOT do that, but that's NOT the way it actually happens. When and if that's the way it actually happens, I'll be happy to change my tune.

      As for your other points, I have no problem with any of it, except that you didn't address my original problem with your characterization of "deficit spending." You instead broke deficit spending into its constituent parts. Deficit spending implies the net effect of BOTH parts! Had you written this:

      "For the banks to own reserves which aren't simply borrowed from the Fed, either the govt has to have previously spent, or the Fed must have purchased assets from the private sector.

      In either case, the government, either through Fed lending or purchases, or through previous spending, has to provide the funds which are then used to pay taxes or to buy Treasury bonds."

      I would have technically* had NO problem with that! But instead of "spending" you wrote "deficit spending" which I do have a problem with.

      *My one caveat is that these sentences still would have been misleading. The way the system works now, not hypothetically, is the following:

      Treasury can only spend what's in the TGA. No more!

      TGA is ONLY** funded (with reserves) by borrowing from or taxing the private sector.

      Reserves can only be obtained from the Fed.

      After Treasury spends, the private sector can repay the Fed.

      Thus reserves must necessarily flow THROUGH the private sector to fund the TGA. You can't cut out the middle man until the rules change.

      **The only exception being interest remitted to Treasury from assets held by the Fed.



      Delete
    27. On the econviz "macroeconomic balance sheet visualizer"

      http://econviz.org/macroeconomic-balance-sheet-visualizer/

      Turn on the 'step-by-step walkthrough' and go to step #6. It says:

      "Government bond issuance comes after government spending and simply gives the private sector a longer duration asset in exchange for money (i.e., it's an asset swap)."

      Delete
    28. phi, I see it now, thanks. Step #9 in this "step by step" process is somewhat telling ("Government Spends (Without Borrowing)"). It says:

      "A sovereign government does not need to 'raise money' before it can spend, and this combined operation demonstrates why."

      But then it also says (as I write above) in the notes for the corresponding pull down operation (outside of step by step mode):

      "Governments currently do not use this spending approach, but they could if desired."

      So you got me there! I don't know why it says that for Step #6... especially when the visualizer doesn't actually obey that same rule! Try running "Government Spends" with $50 (from the initial set of balance sheets). Econviz won't let you do it! It says:

      "Invalid Operation: One or more balance sheets has insufficient assets or liabilities. Try another operation first or reset the balance sheets."

      The only reason you can run it with $40 or less initially is because they start you off with $40 in the TGA!

      But I shouldn't have to convince you of any of this. You already know it! you already granted me that the way we do it now requires there to be funds in TGA prior to spending.

      I think your point was that this was just a choice of the government and the government could choose to let Treasury do "Government Spends (Without Borrowing)" (essentially) at any time without regard to funds in the TGA. Agreed! ... and when that becomes a regular part of Treasury activity, I'll be right there to agree with you. Clearly its not the case now.

      Delete
    29. Tom, I agree that the Treasury is required to have a positive balance in its account at the Fed before it makes payments from that account (i.e. it's not allowed to run overdrafts). I don't know anyone who would disagree with that, by the way. Yes, it's a constraint imposed by the government on both Treasury and Fed operations.

      "After Treasury spends, the private sector can repay the Fed".

      Your description isn't how it works in the US. This 2010 paper by Scott Fullwiler gives a detailed description of US Treasury debt operations, let me know what you think:

      http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1825303

      Delete
    30. phil, I see nothing in Fullwiler's article on first glance to contradict what I said. He concurs that reserves only come from the Fed:

      "Recall that only reserve balances can settle Treasury auctions via Fedwire. Note, though, that the only sources of reserve balances over time (that is, aside from various short-term effects from autonomous changes to the Fed’s balance sheet) are loans from the Fed or the Fed’s purchases of financial assets either outright or in repurchase agreements."

      So, if we start with blank balance sheets for everyone (Treasury, Fed, private sector) then Treasury sells T-bonds, where do you suppose the reserves come from to settle the payment for the T-bonds and fund the TGA? They MUST come from the Fed. The Fed can accept the T-bonds as collateral for the loans, or buy the T-bonds outright, but the reserves come from the Fed!

      Reserves from Fed to private sector to TGA and back to the private sector when Treasury spends. Now the Fed can sell T-bonds or be repaid for their loans in reserves (thus releasing the T-bonds as collateral).

      Is your complaint that I didn't ALSO state that the Fed can buy the T-bonds from the private sector? Scott clearly mentions both channels, and prior to QE, the repo channel was clearly dominant (I don't know the breakdown now). Repos are repaid.

      Delete
    31. "He concurs that reserves only come from the Fed"

      No one disagrees with that. Even if the Treasury minted a trillion dollar coin and deposited it at the Fed, the reserves created by the Treasury's subsequent spending would still be created by the Fed.

      "So, if we start with blank balance sheets for everyone"

      That's not a realistic assumption, the balance sheets are never blank.

      The Treasury always maintains a positive balance in its Fed account (pre QE it was around $5 billion, now I think it's a bit less). The Treasury's account has never hit zero, to my knowledge.

      "where do you suppose the reserves come from to settle the payment for the T-bonds and fund the TGA?"

      When the Treasury spends, this adds reserves to the banking system. These are then drained (if there is an excess) by either the Fed or Treasury selling bonds, or by the Treasury calling in TT&L deposits.

      Prior to/ during settlement of a Treasury bond auction, the Fed adds reserves by buying bonds (created as the result of previous deficits), thereby ensuring that there are sufficient reserves in the system to settle the auction without the fed funds rate rising.

      When the Tsy spends, this would leave the banking system with excess reserves if those reserves were not "drained" in some way. The way they are drained is by either: 1. the Fed selling bonds, 2. the Tsy selling bonds (thereby replenishing the TGA), 3. Tsy calling in TT&L deposits (thereby replenishing the TGA). These operations are coordinated between the Tsy and Fed so as to minimise any impact on the interest rate.

      "The Fed can accept the T-bonds as collateral for the loans"

      T-bonds are required as collateral for loans from the discount window, to my knowledge. For the non-govt to own T-bonds there must have been prior deficit spending.

      So it's a bit of a chicken and egg question!

      Delete
    32. apparently T-bonds aren't the only acceptable collateral for discount window loans:

      http://www.frbdiscountwindow.org/FRcollguidelines.pdf

      Nonetheless, banks have capital requirements.

      I've been thinking a bit about this 'chicken and egg' question, and it inevitably leads back to the system which preceeded the current one (i.e. back to the various forms of gold standard), and then back even further beyond that. i.e long before the Fed even existed...

      We know that the Treasury existed before the Fed. However, trying to account for all of this in a logical sequence from year zero to now is rather difficult...

      Delete
    33. I agree that completely blank balance sheets are probably not realistic, and I agree with the chicken and egg nature you describe. However, I still maintain that true Treasury *deficit* spending is neutral wrt injecting reserves into the private sector. In contrast, Treasury spending (by itself), w/o the "deficit" part ("deficit" implies raising the funds in the TGA through bond auctions), does inject net reserves into the private sector.

      So again, I think we are in almost complete agreement. This is the only sentence I have trouble with: "For the non-govt to own T-bonds there must have been prior deficit spending." Suppose we start with no T-bonds or excess reserves in existence. What's the matter with the following sequence?:

      1. Treas auctions T-bonds to PD, proceeds go to TT&L.
      2. Treas initiates transfer of TT&L to TGA.
      3. PD puts T-bonds up for collateral to borrow reserves from Fed to fund transfer.

      I'm not absolutely clear on the ordering of 2. and 3.... but I don't think it makes too much difference... I realize there's the problem of keeping reserves well balanced to defend the overnight rate. The T-bond is 100% risk free so it contributes nothing to the risk weighted assets in the denominator of the CAR (for calculating capital requirements).

      Prior to 1. the bank could have raised capital and grown its balance sheet in a number of ways by interacting with other private entities (selling stock, making loans, charging fees, etc.... see my Examples 3, 3.1, 3.2 and 7.), if you think that's important. I realize there's the problem of required reserves, but I notice that mortgage backed securities are on your list of acceptable collateral, and since that's for no more than 10% of *demand* deposits (only), it seems plausible to me that MBS would be OK collateral for borrowing that from the Fed. And again, capital can (in principal) be raised up front through stock sales and loan origination fees.

      Conceptually, starting from all blank balance sheets is helpful I think. I don't see anything which would in principle prevent it.

      Thanks for the collateral link BTW.

      Delete
    34. "And again, capital can (in principal) be raised up front through stock sales and loan origination fees." ... or by obtaining certain kinds of liabilities. Here's how John Carney phrases it:

      "To raise the $10 of required capital, Scratch Bank will have to sell shares, raise equity-like debt or retain earnings."

      In this article:

      http://www.cnbc.com/id/100497710

      I asked Joe in Accounting (at pragcap) about the "equity-like debt" part and he mentioned "convertible bonds" I think. I'm still not entirely clear on that, but I think the idea is generally

      measure of capital = set of assets - set of liabilities

      Thus if some liabilities can be converted to something outside this set of liabilities (but still inhabit the right hand side of the BS), they can function to increase your measure of capital.

      I go into some related detail based on an exchange w/ Joe in my Example #7.

      Delete
    35. "I still maintain that true Treasury *deficit* spending is neutral wrt injecting reserves into the private sector"

      The Fed creates reserves, there's no disagreement about that. However in practice Treasury spending increases bank reserve balances. Here's the Fed on the subject:

      "The Federal Reserve is the fiscal agent of the U.S. Treasury. Major outlays of the Treasury are paid from the Treasury's general account at the Federal Reserve.

      The Treasury's receipts and expenditures affect not only the balance the Treasury holds at the Federal Reserve, they also affect the balances in the accounts that depository institutions maintain at the Reserve Banks. When the Treasury makes a payment from its general account, funds flow from that account into the account of a depository institution either for that institution or for one of the institution's customers. As a result, all else equal, a decline in the balances held in the Treasury's general account results in an increase in the deposits of depository institutions. Conversely, funds that flow into the Treasury's account drain balances from the deposits of depository institutions. These changes do not rely on the nature of the transaction. A tax payment to the Treasury's account reduces the deposits of depository institutions in the same way that the transfer of funds does when a private citizen purchases Treasury debt. Both actions result in funds flowing from a depository institution's account into the Treasury's account.

      With the dramatic expansion of the Federal Reserve's liquidity facilities, the Treasury agreed to establish the Supplementary Financing Program (SFP) in order to assist the Federal Reserve in its implementation of monetary policy. Under the SFP, the Treasury issues short-term debt and places the proceeds in the Supplementary Financing Account at the Federal Reserve. When the Treasury increases the balance it holds in this account, the effect is to drain deposits from accounts of depository institutions at the Federal Reserve. In the event, the implementation of the SFP thus helped offset, somewhat, the rapid rise in balances that resulted from the creation and expansion of Federal Reserve liquidity facilities".

      http://www.federalreserve.gov/monetarypolicy/bst_frliabilities.htm


      Delete
    36. "This is the only sentence I have trouble with: "For the non-govt to own T-bonds there must have been prior deficit spending." Suppose we start with no T-bonds or excess reserves in existence. What's the matter with the following sequence?"

      the bank would have to post the bonds as collateral against the TTL deposits, though it could potentially use other forms of collateral to borrow from the Fed.

      Once it has borrowed from the Fed, how does it repay the Fed, in your example? Seems to me the Treasury has to spend, or else the bank has to sell the bonds to the Fed.

      I've emailed the Fed to find out whether banks/PDs can borrow directly from the Fed (discount window) to buy Treasuries at auction, and if they can, whether they do this often or not. OMOs are the primary way in which the Fed adds reserve balances, but it could be that direct borrowing from the Fed (to pay for Treasuries at auction) also happens to some degree. I'll let you know when they get back to me.

      To my knowledge, the Fed only lends directly through the discount window to meet short term needs in unusual circumstances.

      Normally, the Fed buys Treasuries, adding reserves, which are then used to pay for newly-auctioned Treasuries.

      Either way the funds to buy Treasuries come from the government: Fed purchases, fed loans, Treasury spending. Though as you say, the private sector buys the bonds.

      Delete
    37. Yes, I left off #4: Treasury spends... returning reserves back to the banking system which can then be borrowed on the interbank market and used to repay the Fed. I'm not saying that the bank needs to borrow to make the purchase (it does that by crediting the TT&L account out of thin air -- as it does when it makes a loan to a private non-bank)... but instead it needs to borrow from the Fed to clear the transfer from TT&L to TGA. The purchase they make my crediting the TT&L account out of thin air.

      I realize that OMOs are the primary way that the Fed adds reserve balances... that could work too:

      1. Treas auctions T-bond to PD, proceeds go to TT&L.
      2. Treas initiates transfer of TT&L to TGA.
      3. PD sells T-bond to Fed to raise reserves to clear transfer
      4. Treasury spends putting reserves back into the banking system.

      Now could 3. be a repo, such that when the reserves are back in the banking system due to Treasury spending, the bank repurchases the T-bond form the Fed? Is that more likely? I really don't know about the time frame of the various ways that the banks obtain reserves from the Fed to clear the TT&L to TGA transfer.

      Also TT&L accounts could be drained more slowly and from various banks at different times. I've presented it (based on my setup of no bonds in existence) as an all or nothing arrangement... but this could be staged to happen more gradually.

      I'm curious to know how it works, but in the end it doesn't matter too much (to me) if the reserves are obtained through the discount window or Fed repos or Fed purchases. The former two imply that the bank eventually ends up w/ the T-bond again whereas the latter doesn't necessarily (could be the Fed T-bond purchase was permanent).

      I'll be curious to find out what you hear in response to your email.

      I'm supposing here that the banks "own" the T-bond once they credit (again, out of thin air) the TT&L account. That relieves them of having to "borrow to make the T-bond purchase" in the first place. But again, overall, I'm not sure how important that really is.

      Delete
    38. "it does that by crediting the TT&L account out of thin air -- as it does when it makes a loan to a private non-bank"

      TT&L deposits are loans from the Treasury to commercial banks. Banks repay the loans when the Treasury calls in the deposits:

      "note option depositories may hold TT&L balances for an extended time period. As a result, note option banks have use of the Treasury's funds for loans and investments and are required to pay interest to the Treasury. The interest rate is determined by subtracting twenty-five basis points (one-quarter of a percentage point) from the average federal funds rate for the week during which the balances are held".

      http://www.ny.frb.org/aboutthefed/fedpoint/fed21.html

      When a bank credits a TT&L account it's similar to when it makes a payment to a private customer, by crediting their account. In both cases the deposit created represents money owed by the bank.

      Delete
    39. I agree the TT&L account is very similar to how a bank makes a loan to a private customer in the way it credit's the deposit. So as to who is loaning who the money... in both cases (TT&L or private deposits resulting from loans), it's almost better to describe it as an exchange of IOUs.

      In the private entity case, the bank holds the loan papers as an asset. In the TT&L case the bank holds the bond. In both cases the deposit was created ex-nihilo. And in both cases (except for some checking accounts) the banks pay depositors interest.

      Delete
    40. I guess in the TT&L case (unlike the private case) the depositor (Treasury) dictates the amount of interest to be paid on the deposit!

      Delete
    41. The way it normally happens is the Treasury sells bonds, taking reserves from the banking system into the TGA. The Treasury then deposits anything in excess of its normal TGA balance into TT&L accounts at commercial banks. Banks holding these TT&L deposits have to post collateral, generally in the form of bonds, against them.

      To my knowledge banks don't pay for bonds by crediting TT&L accounts. According to Eric Tymoigne it's not current practice, though it has been done at times in the past. It's done at the discretion of either the Fed or Treasury.

      Tax payments, on the other hand, are deposited into short-term TT&L accounts before they are transferred to the Treasury, or else placed into longer-term TT&L accounts held at "note option depositories".

      If a bank were to buy a T-bond by crediting its TT&L account, that would be similar in some respects to someone getting an ordinary loan from a bank. In both cases the deposit created by the bank represents the bank's promise to pay.

      The main difference in the case of the Treasury is that the Treasury always requires payment - in "outside money" (i.e. reserve balances). In other words, it always "withdraws" its deposit. The liabilities of the Treasury's bank - the Fed - are "outside money"; this is what the Treasury requires in payment, and what it uses to spend.

      Delete
    42. I saw the process outlined on econviz, which the author apparently got from Stephanie Kelton. Here's what he writes:

      Explanation of selected operation (Government Issues Debt (Banks Buy via TT&L)):

      "In this example, a commercial bank (known as a Special Depository) buys debt issued by the treasury.

      This is step two of the process outlined by Stephanie Kelton in this blog post. As she notes:
      The treasury sells debt to the commercial bank, and the commercial bank pays for these bonds by crediting the treasury's TT&L (Treasury Tax & Loan) account.
      This is analogous to a bank acquiring a loan by crediting the borrower's account. It does not "cost" the bank anything. No existing resources (reserves) are spent when the bond is acquired.
      The commercial bank gets the bonds (assets) and the treasury gets a credit to its TT&L account at the commercial bank."

      That's from this page:

      http://econviz.org/macroeconomic-balance-sheet-visualizer/

      in the explanation at the bottom for the operation called "Government Issues Debt (Banks Buy via TT&L)"

      Unfortunately the link he gives for Kelton's blog post is broken...

      Delete
    43. Regarding the Treasury always requiring payment in "outside money" ... commercial banks largely do the same! If person x borrows money from Bank A and transfers his deposit to Bank B or purchases goods or services from person y, who does her banking at Bank B, then Bank B will demand payment in "outside money" (reserves). The only exception to this is when person y does her banking at Bank A as well... then no reserves need be transferred. But I'm sure you already knew that. For anybody reading this that doesn't, see my Examples 1, 1.1 and 5.

      Delete
    44. It's an option, but one I think isn't used, or hasn't been used for a while. Either the Treasury or Fed has to decide whether to let banks do it. I'll try and get some more info.

      This FRBNY publication is worth a read. It talks about TT&L deposits.

      http://www.newyorkfed.org/research/current_issues/ci18-3.pdf


      Delete
    45. the above comment was in response to "I saw the process outlined on econviz"

      Delete
    46. "Regarding the Treasury always requiring payment in "outside money" ... commercial banks largely do the same!"

      What I meant is this:

      when you take a loan from a bank, you generally don't ask to withdraw the full amount in cash. If you did, the bank would be loaning you the full amount in cash, instead of just creating a deposit in your account (that deposit being, among other things, a promise to pay you cash on demand).

      In the example of a bank buying a bond by crediting the TTL account, whilst this has similarities to an ordinary loan, a difference is that the full amount is always withdrawn by the Treasury as "cash" (i.e. reserves).

      Commercial banks and the Treasury bank at the Fed. The Fed's deposits are "outside money", so when payment is made from one Fed account to another, it is made in "outside money". What this means in practice is the Fed deletes numbers in one account and types numbers into another.

      Payments between banks are a bit different however, as banks also use private clearing houses to net out their payments, thereby minimizing the amount they have to pay to each other in actual reserves or "outside money". When making payments to the Treasury the only "clearing house" is the Fed.

      In an example above you talked about an imaginary world in which there is only one bank, and how in that world there would be no need for reserves.

      Well, that example is actually kind of real, except that the single bank you're talking about is the Fed. The Fed has no need for settlement reserves, as it's the only central bank in the US. It just creates and deletes deposits, which the rest of the system uses as money.*

      However, if the Fed needs to make payments to foreign central banks (in foreign currency) then it needs to have reserves of foreign currency, as it can't just create them out of thin air.

      So to the Fed foreign currency is "outside money", whilst any dollar-denominated money is "inside money" (including its own deposits and notes).

      "Inside" and "outside" are relative terms, not absolutes.

      Perry Mehrling talks about this understanding of inside and outside money in the paper I linked to. Mosler has also discussed it in blog posts.

      *(The assets that "back" the Fed's deposits and notes are mainly Treasury bonds).

      Delete
    47. "Payments between banks are a bit different however, as banks also use private clearing houses to net out their payments, thereby minimizing the amount they have to pay to each other in actual reserves or "outside money"."

      You're saying that on average there's going to be some funds flowing from bank A to bank B and some from bank B to bank A, so the total amount of reserves transferred can be minimized, but the net payment is still demanded in reserves!

      Similarly, some customers will be withdrawing their loans in cash, or partly in cash, and some customers will be depositing cash... so the net flow of cash out of the bank can be minimized.

      The difference I see w/ Treasury bond purchases, is that funds tend to flow in one direction (banks to Treasury) so there isn't much "netting out" for the Fed to do.

      The principle of how reserves flow between deposit holders is the same.

      Delete
    48. BTW, if anybody reading this exchange is confused, I recommend this Scott Fullwiler article:

      http://www.nakedcapitalism.com/2012/04/scott-fullwiler-krugmans-flashing-neon-sign.html

      Especially see figure 2.

      Delete
    49. "the net payment is still demanded in reserves!"

      Agreed. For example, If I owe you $100, and you owe me $101, we can settle our accounts by you giving me $1. The debts we owe each other could be described as "inside money". The money we use to settle those debts could be described as "outside money".

      "the net flow of cash out of the bank can be minimized".

      Yes. Banks try to keep as much money as possible "in house". They generally do this by offering their services and by paying interest on deposits.

      "funds tend to flow in one direction (banks to Treasury)"

      When the Treasury spends, funds flow in the other direction, no?

      If these funds aren't used to buy Treasuries, or else withdrawn through calls on TTL accounts, then - in reality - banks end up with excess reserves.

      In theory it's possible that they might not, but in practice they do.

      As it says in that FRBNY publication:

      "Flows of funds between the TGA and private depository
      institutions were important prior to the crisis because the TGA is maintained on the books of the Federal Reserve; increases in TGA balances stemming from Treasury net receipts drained reserves from the banking system and, in the absence of offsetting actions, put upward pressure on the federal funds rate. Conversely, decreases in TGA balances resulting from Treasury net expenditures added reserves to the banking system and, absent offsetting actions, put downward pressure on the funds rate".

      "The principle of how reserves flow between deposit holders is the same".

      If we both have accounts at bank A, and you pay me $100, then bank A debits $100 from your account and credits $100 to my account. No reserve or cash transfer is necessary, as you have mentioned before. Our bank simply shifts its debts from you to me.

      If, however, we both have accounts at the Fed, and I'm the Treasury and you're a bank, then payment necessarily involves a transfer of reserve balances from you to me. In other words the Fed debits your reserve account and credits my account. This is because the fed's deposits are "reserves". Agreed?



      Delete
    50. *"They generally do this by offering their services and by paying interest on deposits".

      Also, loans "bring deposits back into the bank" via repayment of interest and principal.



      Delete
    51. "When the Treasury spends, funds flow in the other direction, no?"

      Sure, but that's why I wrote "The difference I see w/ Treasury bond purchases" before that.

      I didn't mean to include the WHOLE picture... but now that you mention it, I wonder if the Fed does do a bit of that. Say at the end of each day they might say:

      Treasury pays person x $100, and person x has his account at bank A, and Treasury is also moving $50 from the TT&L at bank A to the TGA... so that's net $50 of reserves from TGA to bank A. Do you suppose the Fed does a bit of "netting out" at the end of each day along these lines?

      Delete
    52. "If, however, we both have accounts at the Fed, and I'm the Treasury and you're a bank, then payment necessarily involves a transfer of reserve balances from you to me. In other words the Fed debits your reserve account and credits my account. This is because the fed's deposits are "reserves". Agreed?"

      Agreed. But the analogy still holds.

      Case 1: persons x and y have deposit accounts at bank A, and x pays y. Then bank A simply debits x's deposit and credit's y's. An example of a pure inside money flow.

      Case 2: entities x and y have deposit accounts at the Fed, and x pays y. Then the Fed simply debits x's deposit and credit's y's. An example of a pure outside money flow.

      Now why did I put "pure" in there in both cases? Originally I didn't have it, but decided to put it in, because some flows involve a mix. For example, in Case 1, when x and y have deposits at separate banks, or in Case 2, when x is Treasury and y is a commercial bank, and x is transferring reserves to y to back a payment to a 3rd party, say person z, who's deposit is at y. In this case z's (inside money) deposit is also credited.

      Delete
    53. "Do you suppose the Fed does a bit of "netting out" at the end of each day along these lines?"

      I don't think the Fed "nets" payments between the Treasury and banks. My guess is that it doesn't, as Fedwire is a "real-time" gross settlement system, where every individual payment is instant, final and irrevocable.

      There is no equivalent to a private "clearing house" system in which payments are gathered up and then "netted out" by the end of the day.

      Delete
    54. sorry, there wasn't any need for all those "speech marks".

      Fedwire: http://en.wikipedia.org/wiki/Fedwire

      Delete
    55. "An example of a pure inside money flow"

      I think it might be a mistake to talk about money as if it’s a sort of commodity, which gets produced and then passed around.

      In the current system money is basically just accounting entries. These are generally held in the form of digital records, or else on pieces of paper.


      When the Treasury spends, it often sends out a type of ‘accounting entry’ on a piece of paper, called a check. This check represents a promise to pay from the Treasury to the person receiving it.

      The person receiving this check then usually deposits it at their bank.

      The bank then types numbers into their account (i.e. the bank credits their account).

      These numbers or credits represent promises to pay from the bank to the depositor.

      The bank then sends the check on to the Fed for payment.

      When the Fed recieves the check it types numbers into the bank's reserve account (i.e. the Fed credits the bank's account), and debits the Treasury's account.


      The thing which gets passed around in this example is the check - a record of the Treasury’s promise to pay.

      The transfer of money, however, is simply an act of changing numbers on different balance sheets.

      The Treasury doesn't literally put money into the person's bank account. Rather, the bank types numbers into their account - i.e. credits their account.

      In turn, the Fed types numbers into the bank's reserve account - credits their account - when the bank 'deposits' the check at the Fed.

      There’s no actual “inside” or “outside” money flow in this example. There are simply different types of promises and payments being made by different types of organizations.

      Delete
    56. phil, of course the "flow" doesn't necessarily involve notes and coins or physical objects. The accounting entries ultimately document the "flow" I speak of. The accounting is constrained to follow rules which is consistent with a "flow of funds." I'd say the same thing if we had an all electronic system with no coins or paper checks or bills of any kind.

      Delete
    57. Ok.

      There is a tendency over at Pragcap to talk about money like it's some sort of commodity that gets produced by banks and then passed around.

      For example, it's not logically possible for the Treasury to "obtain your bank deposit and then spend it", yet this sort of thing is frequently written over there, for some reason.

      Delete
  3. 2) "Federal government deficit spending does NOT inject reserves into the banking system"

    This seems to contradict your earlier statement:

    "Here, then, is the list of ways in which reserves leave the private banking system...

    2. When Treasury auctions bonds (and the proceeds are transferred to the TGA)"

    If reserves leave the banking system when the treasury sells bonds, then logically reserves must enter the banking system when the government deficit spends.

    Pre-QE (when banks did not hold ‘excess’ reserves and instead lent them to each other at a positive rate of interest) government deficit spending would create excess reserves within the banking system, meaning that the Fed would have to sell govt bonds to banks (to drain reserves) if it wanted to maintain its interest rate target.

    3) "Thus the Federal government obtains and spends private bank created money (inside money) when it deficit spends."

    How? Can you do a balance sheet example which shows this?

    Private bank liabilities cannot 'go into' the TGA. The TGA cannot 'obtain' private bank liabilities. But as we know the government spends from the TGA. So, logically, the government cannot "obtain and spend" private bank liabilities in the way that you suggest.

    The government does however obtain bank liabilities when its TT&L accounts are credited. These bank liabilities are loans from the Treasury to the banks at which the TT&L accounts are held. When the Treasury calls in these TT&L deposits, the bank's liability is extinguished and its reserve account is debited. Payment to and from the TGA is made with reserves - i.e. Fed liabilities - not with private bank liabilities.

    Furthermore, the idea that a bank takes someone's deposit and puts it into the TT&L account doesn't really make much sense. Bank deposits are basically debts owed by a bank to its depositors.

    Say Mr.A pays $100 tax. His bank debits his account and credits the TT&L account (before payment is made by the bank to the TGA at a later time or date). The bank now no longer owes Mr.A $100. Instead it now owes the Treasury $100. It makes little sense to describe this process as 'taking Mr.A's deposit and giving it to the government'. The bank's debt to Mr.A (Mr. A’s deposit) has simply been extinguished.

    Discuss?

    ReplyDelete
  4. phil, you write:

    "If reserves leave the banking system when the treasury sells bonds, then logically reserves must enter the banking system when the government deficit spends."

    Yes, however in this context "deficit spending" is meant to encompass the whole process 1) Treas auctions bonds 2) Gov spends proceeds from TGA. I'm skipping over the TT&L account business, but you can assume that's part of getting the funds into the TGA account. I'm also assuming all the proceeds are spent. My argument here is exactly parallel to the operation entitled "Government Spends (Consolidated)" on this website:

    http://econviz.org/macroeconomic-balance-sheet-visualizer/

    Sorry that wasn't clear, but I thought it was implied by "deficit" which implies the bond sales.

    You write:

    "Pre-QE (when banks did not hold ‘excess’ reserves and instead lent them to each other at a positive rate of interest) government deficit spending would create excess reserves within the banking system, meaning that the Fed would have to sell govt bonds to banks (to drain reserves) if it wanted to maintain its interest rate target."

    Again, if by deficit spending we include the whole process from selling the Treasury bonds to spending all the proceeds of the bond sales, that's not true. I agree that in isolation government spending (of any sort) puts reserves back into the banking system.

    You write:

    "How? Can you do a balance sheet example which shows this?" Yes, I think I can. But such an example already exists on the econviz.org website. Again it's the "Consolidated" spending. You can do all the operations of the "Consolidated" spending individually there as well. As far as I can tell, everything there is correct.

    Re: your comments on TT&L accounts. Yes, I agree that I'm skipping over the mechanics of how these accounts fit into the picture. Again the econviz site breaks each step out in isolation. I consider those to be unimportant details for the big picture however. To create an example around your comment (I'm changing Mr. A to Mr. X to keep with my usual naming conventions):

    Say Mr. X borrows $100 from Bank A to pay his tax with.

    Fed and Treasury: blank balance sheets

    Bank A:
    Assets: $100 loan to x
    Liabilities: $100 deposit for x

    Person x:
    Assets: $100 deposit at A
    Liabilities: $100 borrowing from A

    Now after the tax has been paid and the TT&L account has been credited and the funds transferred to the TGA account we have:

    Fed:
    Assets: $100 loan to A
    Liabilities: $100 reserves in TGA

    TGA:
    Assets: $100 reserves
    Liabilities: $0

    Bank A:
    Assets: $100 loan to x
    Liabilities: $100 borrowing from Fed

    Person x:
    Assets: $0
    Liabilities: $100 borrowing from A

    Now suppose the gov spends all $100 on Person x's services, and Bank A uses the reserves (transferred from the TGA as part of the gov's payment to x) to pay back its loan from the Fed. Now we're back to our original set of balance sheets. The Fed and TGA have clear sheets and it looks exactly like Person x just took a loan from Bank A.

    I'm using "Person x" as a stand in for all non-bank individuals and private businesses and I'm using "Bank A" as a stand in for all private banks.

    Now this example was NOT deficit spending, but we could do a similar set for deficit spending in which case Person x would end up with a $100 T-bond. See my Example 6 for a equally over-simplified example of deficit spending. Think of the "bond auction" as a Treasury direct sale, in which case I don't think the TT&L accounts come into play. That's a nice way to avoid those details. Perhaps I should demonstrate those details in an example, but I wouldn't be adding anything that econviz hasn't already covered.











    ReplyDelete
    Replies
    1. The TT&L accounts aren't necessary in a basic description, as you say they are an intermediate step.

      "Again, if by deficit spending we include the whole process from selling the Treasury bonds to spending all the proceeds of the bond sales, that's not true".

      Pre-QE banks didn't hold excess reserves. Which means that they didn't have any excess reserves with which to buy government bonds. If banks were to use whatever reserves they had to buy government bonds, this would raise the interbank interest rate as reserves left the banking system by going into the TGA. So the Fed had to provide the reserves to settle bond auctions if it wanted to hit its target rate. The way it generally did this is by buying bonds from banks (usually with a repo agreement). This provided the banks with the needed reserves to purchase the govt bonds.

      Then, at a later time, when the Treasury spent, the banking system would get excess reserves. So the Fed would have to sell bonds to the banks to drain those reserves, if it wanted to hit its target interest rate. In the case of a repo agreement, this would be pre-arranged.

      Delete
    2. I say "pre-QE" because things are a bit different now, given that banks hold a massive amount of excess reserves.

      Delete
    3. In your example above, Mr.X is currently in debt to bank A and bank A is in debt to the Fed. As it stands, neither Mr.X nor bank A can repay their debts.

      But this is somewhat besides the point as nowhere in your example have you shown that the Treasury has obtained or spent private bank deposits. It has simply obtained a deposit at the Fed.

      Delete
  5. phil, you write:

    "The TT&L accounts..."

    Yes, I agree, but at the end of this process you state the Fed sells bonds back to the banks to drain the reserves to hit its interest rate target. Yes I skipped over that bit too in the interests of simplicity. I just depicted this as the Fed loaning the reserves out. But at the conclusion of this process I think you and I agree: no excess reserves.

    You write:

    "I say "pre-QE" .... "

    Yes, but the result is the same. Once the spending has been done, no net excess reserves have been added. Some were taken away temporarily but then spent right back into the banking system by the gov.

    "In your example above..."

    True, Treasury has not spent private bank deposits in a very narrow technical way of seeing it. But I still claim that the vast majority of money used to fund the treasury ultimately originated from private loans. If the Fed held no Treasury bonds, then ALL of it could be traced to private loans. I understand the Fed has a role here in either case in loaning reserves to facilitate the process, but those reserves are repaid at the end of the process. I think this is a forest for the trees disagreement.

    ReplyDelete
    Replies
    1. Keep in mind that the majority of Treasury bonds are held by the investing public, and thus the banks are really acting as intermediaries most of the time. That's important to point out. That public can't borrow reserves from the Fed to finance their bond purchases... they had to acquire their funds (again, ~97.5% of which originated from private loans to private individuals) with which to purchase the bonds through other means. So what I'm saying is that when private non-bank investors purchase bonds the bulk of the money they're using originated from private bank loans to private non-bank entities. A lot of technical details happen between that purchase and when the government ultimately spends the money, but in the end the composition of the money spent by the government back into the economy doesn't change: it can be traced (mostly) to private loans. If the Fed doesn't hold any Treasuries permanently, then ALL of it can be said to have originated from private loans. This is a little bit like a chicken and the egg kind of argument I think. You're pointing out technical details, which are good to point out... but ultimately are factored out of the equation (from my point of view anyway). At the end of my example, the Fed and gov balance sheets are clear again. There's $100 of deposit money in the system and $100 of a private bank loan from whence it came. You can divide the example up into more actors (add more people and more private banks), but the conclusion would be the same. Some people would be holding a total of $100 in bank deposits, and some people would be borrowing a total of $100 from the banks. If all money is based on debt, then its pretty clear what debts the money came from. Now if the gov runs a surplus, then yes, money is getting sucked out of the system and accumulating in the TGA. And if the gov runs a deficit, then Treasury bonds are accumulating with the public. Ultimately though I view the Fed as providing a facilitating role. I think the natural place to "sample" this process is when the Fed's balance sheet is empty! You could choose to sample it other places (as I gather you do), but to me that doesn't provide much clarity about what is ultimately going on.

      Delete
    2. "True, Treasury has not spent private bank deposits in a very narrow technical way of seeing it".

      Leaving aside the technical description, it seems to me that the idea of the government 'obtaining and spending' bank deposits doesn't make much sense in a general sense either (though you may disagree).

      Bank deposits (liabilities) are promises to pay government/ central bank liabilities (cash/coin/reserves).

      So if the government were to "spend a bank deposit" it would be spending a promise to pay its own liability. Similarly, if the government "obtains a bank deposit" it is obtaining a promise to pay its own liability.

      This is like you giving someone a promise to pay your own IOU, or "obtaining" a promise to pay you your own IOU.

      Delete
    3. "So what I'm saying is that when private non-bank investors purchase bonds the bulk of the money they're using originated from private bank loans to private non-bank entities".

      The point I would make is that government bonds can't actually be purchased with bank deposits. The money paid into the TGA can only be a liability of the Fed.

      The movement of bank liabilities within the private sector is relevant to how different private sector agents finance their own positions, but it doesn't ultimately have anything to do with how the Treasury finances its own operations. The Treasury basically receives and spends Fed liabilities.

      Delete
    4. "The point I would make is that government bonds can't actually be purchased with bank deposits. The money paid into the TGA can only be a liability of the Fed."

      Agreed. But I still think that's missing the bigger picture. I see the Fed's role as that of an intermediary between bank created money and the TGA. The Fed reverses this intermediary role when the gov buys goods and services from the private sector via the TGA account: it not only credits the payee's bank's reserve account, it also instructs the bank to credit the payee with a bank deposit. The exact mirror image of how money flows into the TGA (via taxes). And as you point out, that's exactly when those reserves will be used to repay the Fed's loans to the banks as well. So to summarize a complete cycle:

      Money flows into TGA from private sector (by taxes):

      bank deposit goes down
      Fed loan of reserves to banks goes up
      TGA account goes up

      Money flows out of TGA back into private sector:

      TGA account goes down
      bank deposit goes up
      Fed loan of reserves to banks gets repaid

      The Fed's hand is forced in this process. It's the gov that taxes and spends.

      Now if it's deficit spending we're talking about, there's an added twist:

      Money raised by Treas bond sale goes into TGA account:

      bank deposit goes down but bond holdings go up
      Fed loan of reserves to bank goes up
      TGA account goes up

      The spending process is the same (see above), so overall the only difference is the private sector is left with a net financial asset (NFA) in the end, i.e. the T-bond. Otherwise the process is the same as through tax and spend. The Fed's intermediary role is identical, and they are again forced into it. It's the gov that decides to sell the bond and spend the proceeds, and it's the private sector which decides to buy the bond.

      Delete
  6. Hi Tom, thanks for pointing me to your blog from Cullen's page.

    I don't know if you've seen it, but this publication from the NY Fed is what really helped me to understand how reserves work: http://www.newyorkfed.org/research/current_issues/ci15-8.pdf

    What I can't get my head around is how excess reserves would ever leave the private banking system and cause inflation. I am thinking of inflation in terms of the general public having more money to spend, which would cause prices to rise and the value of the dollar to decrease. How does the public get the money in their hands?

    Consider:

    The only way for money to exit the private system is to either go back to the Fed/Treasury (1,2,4 and 6 in your example, although 1, 2 and 6 are recycled back into the private system, i.e. redistributing to Paul); to be withdrawn in the form of hard currency (minimal); or to be transferred to an international bank.

    Is this correct? If so, (and ignoring hard currency withdrawals), how could excess cash from QE ever cause inflation? I mean, presuming that the excess cash all stays inside the larger banking system (only moving from bank to bank)? The Fed is paying IOER, which means the banks don't loan their excess reserves to each other, which means interet rates can't go below the floor set by the Fed. The Fed could also sterlize (mop up) the excess money by selling assets and then.....just remove the proceeds from the system (? I assume they can do this).

    I have done a ton of reading on this and this is the last bit that I'm not clear on. You, Cullen and others have well established that excess money does not necessarily lead to inflation (because it's not "new" money going into the system, but an asset swap), but how COULD it lead to inflation? What would be the mechanism for this happening? This is driving me nuts.

    ReplyDelete
  7. NWH, thanks for the link and the comment. That's an EXCELLENT question you pose! Yes, I think that excess reserves don't necessarily cause inflation. A lot of people thought they would, but they've largely been wrong. Peter Schiff for example has been famous for predicting year after year that we'd have high inflation or hyperinflation. Even at least one of his fellow Austrians has ridiculed him for that (Mish Shedlock). Some people have pointed out that QE is like "pushing on a string" regarding getting anything useful, or anything at all for that matter, to happen.

    I'm REALLY not an expert, so I'm flattered that you thought to ask me, so I'm just telling you my views here. But here's another take: QE in the US has been just an asset swap, however it's true that INDIVIDUAL banks can use those reserves to purchase other assets (if they maintain Good CAMELS scores, etc.) because QE grows there balance sheets. Additionally the former non-bank T-bond holders will have more deposits after QE. The Market Monetarists claim this causes ALL financial asset prices to rise, and the risk adjusted rate of return of financial assets to decrease, thus making it marginally more likely that investors will put their money in something other than financial assets. Perhaps build that new factor or buy a Ferrari. A guy on Scott Sumner's site took some time to explain this view to me. Here you are:

    http://www.themoneyillusion.com/?p=19141#comment-225110

    I think Cullen states that QE (as practiced in the US) can cause "distortions" in the market. I personally would LOVE a concrete example of that!... but I don't really have one.

    QE in Japan, however, is a different beast. There the BOJ is actually purchasing a lot more than government bonds I understand. Even the Market Monetarists here (e.g. Scott Sumner) say the Fed has the ability to "buy the whole world" if need be (which would CLEARLY cause inflation! Ha!), but that it would never need to because it can threaten to (by saying they'll do what's required to raise inflation) and that will suffice.

    "The Fed could also sterlize (mop up) the excess money by selling assets and then.....just remove the proceeds from the system (? I assume they can do this). "

    Yes, that is OMS (open market sales). So rather than being transferred to an international bank (a bank which as a Fed reserve account), I think OMSs are the MAIN way in which reserves leave the banking system "permanently."

    So... ultimately I don't really have a good answer for you! Sorry! Let me know if you ever find a satisfying answer!

    ReplyDelete
  8. Tom, you say above:

    "Only to the extent that the Fed holds Treasury bonds as assets can you say that private bank deposits in the private sector are backed by "loans to the government" (Treasury bonds) rather than "loans to private entities."

    I think this is incorrect.

    Let's use your example of a bank borrowing from the Fed to buy Treasuries (no reply from the Fed on this subject as yet!):

    Initial setup:

    Tsy: 0, Fed: 0, Banks: 0, Non-banks: 0

    1. Banks borrow 100 from the Fed, and buy 100 bonds from the Treasury:

    Tsy: A: 100 fed deposits, L: 100 bonds, E: 0
    Fed: A: 100 loans, L: 100 tsy deposits, E: 0
    Banks: A: 100 bonds, L: 100 loans, E: 0
    Non-banks: A: 0, L: 0, E: 0

    2. Treasury spends 100:

    Tsy: A: 0, L: 100 bonds, NE: 100
    Fed: A: 100 loans, L: 100 reserves, E: 0
    Banks: A: 100 reserves, 100 bonds, L: 100 deposits, 100 loans, E: 0
    Non-banks: A: 100 deposits, L: 0, E: 100

    3. Banks repay 100 loans. Fed buys 10 bonds from banks to meet reserve requirements:

    Tsy: A: 0, L: 100 bonds, NE: 100
    Fed: A: 10 bonds, L: 10 reserves, E: 0
    Banks: A: 10 reserves, 90 bonds, L: 100 deposits, E: 0
    Non-banks: A: 100 deposits, L: 0, E: 100

    You can see here that the amount of deposits in the private sector created by government spending, or "backed by loans to the government", is equal to the quantity of bonds + reserves held by banks.

    4. Let's say now that non-banks buy 90 bonds from banks:

    Tsy: A: 0, L: 100 bonds, NE: 100
    Fed: A: 10 bonds, L: 10 reserves, E: 0
    Banks: A: 10 reserves, L: 10 deposits, E: 0
    Non-banks: A: 10 deposits, 90 bonds, L: 0, E: 100

    Non-bank deposits = bank reserves

    5. Fed sells 9 bonds to banks to drain excess reserves:

    Tsy: A: 0, L: 100 bonds, NE: 100
    Fed: A: 1 bonds, L: 1 reserves, E: 0
    Banks: A: 1 reserves, 9 bonds, L: 10 deposits, E: 0
    Non-banks: A: 10 deposits, 90 bonds, L: 0, E: 100

    Non-bank deposits = bonds held by banks + reserves

    6. Now let's say the Fed does QE, buying 90 bonds from non-banks:

    Tsy: A: 0, L: 100 bonds, NE: 100
    Fed: A: 91 bonds, L: 91 reserves, E: 0
    Banks: A: 91 reserves, 9 bonds, L: 100 deposits, E: 0
    Non-banks: A: 100 deposits, L: 0, E: 100

    Non-bank deposits = bonds held by banks + reserves


    ReplyDelete
  9. phil, there well may be a problem w/ my statement. I didn't mean for it to hold exactly, and I wasn't taking into account reserve requirements and many other details. Perhaps I can clean that up a bit! So, honestly, thanks for looking into it!! I appreciate it. I went through a similar set of example balance sheets myself to come up w/ a general rule, but mine was:

    Treas assets + bank reserves + non-bank cash = Fed liabilities

    Which should have been obvious to me from the start!

    However, originally I also had on the left "non-bank deposit-assets" and on the right "bank-deposit-liabilities," but of course those are the same (not counting retained earnings here from loan origination fees or interest payments, etc.). So I simply subtracted those from both sides to simplify. I also left out foreign CBs, the IMF, GSEs etc.

    So now let me address what you have... I'll check it later in detail for any errors, but I'll assume your math is correct for now.

    Re: T-bonds held by banks, I think those are in general a small % compared to those held by non-banks. Am I correct? Somewhere I read that banks hold about 2% of all the T-bonds out there in private hands... but I couldn't even begin to tell you where I saw that, so I very well may be in error there. But that was my assumption.

    Let's assume for the moment that is correct, then your formula could be simplified to the following approximation:

    non-bank deposits = reserves

    OK, fine. The thing that you're leaving out though, and which I believe cannot be approximated away, are private loans! Humor me for a moment and assume nobody keeps a demand deposit (they're all savings deposits) or that we're in Canada w/ no reserve requirements... then those get added directly to both sides, and we end up with:

    non-bank deposits = reserves + bank loan-assets

    The left hand side is now a combination of what was there previously (non-bank deposits = reserves = T-bonds held by the Fed) + an added non-bank deposit term due to the private loans.

    The reserves in this case representing exclusively T-bond purchases from the Fed!

    Do you see where I'm going with this? Nowhere in your example do the non-banks simply take a loan from the banks! That seems to me a fairly major omission! Granted I made a lot of simplifications here, but I think they're valid.

    That led me to my original statement. Now we can add in the bank held T-bonds and the reserve requirements and the retained earnings and the cash and Treas assets (i.e. unspent funds in the TGA), etc. and come up w/ a more accurate description, but to first order I think my approximation might hold water. Your thoughts?

    ReplyDelete
    Replies
    1. Obviously, if the Treas were to start accumulating a lot of funds in the TGA (i.e. running a surplus), then my approximation is out the window!... or if banks were to hold a large % of privately held T-bonds (actually that was my motivation for making my private set of example balance sheets originally... to try and account for bank assets... which I never really found a satisfactory way to do!).

      Delete
    2. I didn't include private loans because they weren't relevant to the example. The example isn't meant as a model of the economy. It simply serves to demonstrate that your statement, (that "only to the extent that the Fed holds Treasury bonds as assets can you say that private bank deposits in the private sector are backed by "loans to the government") is incorrect.

      Delete
    3. OK, sure it's an approximation rather than absolutely true. My purpose was to divide up private deposits into private debt created vs public debt created... and the public part traces it's linage to the Fed's holdings... not the full Treasury debt. I'll grant you the approximation breaks down if certain conditions change.

      Delete
    4. If you look at the example you can see that government spending results in a net increase in non-bank deposits equal to the amount spent. Govt spends 100, non-bank deposits increase by 100 (net).

      The spending provides income to the non-banks which they can then use to buy government bonds from the banks, if they want to.

      Here's it is again, a bit different this time:

      1. Banks borrow 100 from the Fed, and buy 100 bonds from the Treasury:

      Tsy: A: 100 fed deposit, L: 100 bonds, E: 0
      Fed: A: 100 loans, L: 100 tsy deposit, E: 0
      Banks: A: 100 bonds, L: 100 loans, E: 0

      Peter: A: 0, L: 0, E: 0
      Paul: A: 0, L: 0, E: 0

      Private sector net financial assets are unchanged - bond issuance has not increased private sector NFA.

      2. Treasury spends 100 (pays Paul 100):

      Tsy: A: 0, L: 100 bonds, NE: 100
      Fed: A: 100 loans, L: 100 reserves, E: 0
      Banks: A: 100 reserves, 100 bonds, L: 100 deposit, 100 loans, E: 0

      Peter: A: 0, L: 0, E: 0
      Paul: A: 100 deposit, L: 0, E: 100

      Private sector net financial assets have increased by 100. Reserves have increased by 100. Deposits have increased by 100 (net).

      Paul now has net financial assets in the form of a bank deposit.

      The private sector as a whole has net financial assets in the form of government bonds.

      3. Banks repay 100 loans:

      Tsy: A: 0, L: 100 bonds, NE: 100
      Fed: A: 0, L: 0, E: 0
      Banks: A: 100 bonds, L: 100 deposit, E: 0

      Peter: A: 0, L: 0, E: 0
      Paul: A: 100 deposit, L: 0, E: 100

      Private sector net financial assets are unchanged.

      4. Paul pays Peter 100:

      Tsy: A: 0, L: 100 bonds, NE: 100
      Fed: A: 0, L: 0, E: 0
      Banks: A: 100 bonds, L: 100 deposit, E: 0

      Peter: A: 100 deposit, L: 0, E: 100
      Paul: A: 0, L: 0, E: 0

      Peter now has net financial assets in the form of a bank deposit. Private sector net financial assets are unchanged.

      5. Peter decides to save, so buys 80 bonds from banks:

      Tsy: A: 0, L: 100 bonds, NE: 100
      Fed: A: 0, L: 0, E: 0
      Banks: A: 20 bonds, L: 20 deposit, E: 0

      Peter: A: 20 deposit, 80 bonds, L: 0, E: 100
      Paul: A: 0, L: 0, E: 0

      Private sector net financial assets are unchanged.

      As you can see above there is no "taking from Peter to pay Paul".

      Delete
    5. Here's a list of the primary dealers, by the way:

      http://www.newyorkfed.org/markets/pridealers_current.html

      Delete
    6. Now let's say instead that Peter already has a 100 deposit (originally created by a previous bank loan to someone else), and he decides save and buy 100 bonds from banks at the same time that the govt spends 100.

      In this case deposits would decrease by 100 when Peter buys the bonds, and increase by 100 when the government spends (to pay Paul).

      So, in this case there would no overall increase in deposits.

      However, whereas before there was a 100 deposit created by a private loan, now there is a 100 deposit created by government spending.

      So we could say in this case that the quantity of deposits "backed by loans to the government" equals the quantity of bonds held by Peter.

      However, as before, nothing has been "taken from Peter to pay Paul". The Treasury has not "obtained and spent inside money".

      Note that in this example Peter's net financial assets would be unchanged when he buys the bonds. He would simply be swapping his net financial assets in the form of a bank deposit for net financial assets in the form of bonds.

      Conversely, Paul would experience an increase in net financial assets as a result of the government spending. These net financial assets would be in the form of a bank deposit created by the government spending.

      So it would be completely incorrect to describe this as "the government taking from Peter to pay Paul and giving Peter a NFA in the form of a T-bond".


      Delete
    7. Thanks phil, everything looks fine on first glance, and as I've explained before, I have no problem with saying that Treasury "spending results in a net increase in non-bank deposits equal to the amount spent." It's when you bring in "deficit spending" that I have a problem, but I guess that's because I don't imagine the banks as being the primary holders of T-bonds in the private sector. You've demonstrated that to the extent that banks ARE the bond holders, then net non-bank deposits result from the "deficit spending" process. If they are insignificant holders (i.e. Peter buys $100 of bonds from the bank in your last step), then non-bank deposits do not result. I believe that non-banks ARE the primary bond holders however. However, I'll grant you a point because it's something I've just been assuming (that non-banks are the bond holders) w/o stating it outright.

      Now as to your last line:

      "As you can see above there is no "taking from Peter to pay Paul"."

      You are so right about that... what you've demonstrated instead is taking from the banks to pay Paul! All you did was replace Peter with the banks by making them the primary bold holders when Paul was paid! Nice try though.

      All you have to do is start off with Peter buying the bond instead of the bank, and you have your "Peter to Paul" scenario, and like I assert above, I think that's more typical. Peter may use a bank as an intermediary to obtain the bond, but he doesn't even have to do that. For example, he could buy through "Treasury Direct."

      For a complete example, say we start with Paul taking a loan from the bank to pay Peter for a house, then Peter in turn buys a bond from Treasury Direct, the bank making good on the deposit transfer by borrowing from the Fed with the loan to Paul as collateral. Now Treasury spends the proceeds of the bond auction on work that Paul does for the Def. Dept. Now the bank can repay the Fed w/ reserves, and Paul can pay off his bank loan.

      Am I saying that Treasury get's Peter's bank deposit and turns it around and spends it on Paul? No, **not directly**! However, there is an interface "flow" on either side of Treasury in which events in commercial bank deposits are linked with events in Fed deposits.

      1. When Peter buys the bond through Treas Direct his bank deposit is debited (erased) and Treasury's Fed deposit is credited (through a transfer of the bank's Fed deposit to Treasury, the bank borrowing funds from the Fed to accomplish this: are any of these details Peter's problem? No!).

      2. When Treasury pays Paul, its Fed deposit is transferred to the bank's Fed deposit and Paul's bank-deposit is likewise credited.

      1. and 2. serve as an interface between the Fed deposit world and the bank deposit world. You may not like my terminology and use of the word "flow" or "interface" here, but I don't think there's anything wrong with this example as I've sketched it.

      Did I cherry pick my example to demonstrate deficit spending resulting in funds moved from Peter to pay Paul with Peter being issued a T-bond in the process? Absolutely! Did you cherry pick your example to squeeze Peter out of the picture and let the bank take his place in this process? Absolutely! Did you add Peter back in your example to add confusion! Absolutely! :) (sorry, I couldn't resist that last bit... but it was confusing!)

      Delete
    8. In my example, you might complain that the bank using its mortgage for Paul as collateral to borrow reserves from the Fed is unrealistic since the mortgage is a riskier asset than a T-bond. OK, fine. Assume the mortgage was for $100 and the T-bond is only $50 then (i.e. assume there's a discount involved). MBS is listed on that Fed link you provided a while back as an acceptable collateral.

      Delete
    9. As to your post starting "Now let's say instead that Peter"... (which you must have posted while I was still replying to your earlier post):

      You write:

      "So we could say in this case that the quantity of deposits "backed by loans to the government" equals the quantity of bonds held by Peter."

      Yes, and it also equals the quantity of "previous bank loan to someone else."

      Now say that $100 that Treas spends to pay Paul is invested (by Paul) in another bond issued by Treas, and Treas spends that $100, again on Paul. Hmmm, now what? We've got $200 in bonds, but still just $100 in Paul's account. Repeat eight times... the amount of the loan to Treas keeps climbing, by $100 each time up to $1000, but Paul's deposit is still just $100 at the end of each cycle. Does it make more sense to speak of that $100 as "backed by loans to the government" or as being backed by that original $100 private loan to "somebody else" (which has not been repaid)? I claim the latter, because those figures will always match. Now let's say Paul instead uses his $100 to buy services from that "somebody else" and they in turn pay off their bank loan with it. What now? We've got $1000 in loans to the gov, backing what though? There are no bank deposits! Instead the bank deposits match the mount of exiting private loans: $0.

      You write:

      "However, as before, nothing has been "taken from Peter to pay Paul". The Treasury has not "obtained and spent inside money"."

      See my example above. All you've done is replaced Peter with the commercial banks when Paul was paid.

      Delete
    10. Now if Stephanie Kelton is correct (assuming econviz correctly interpreted her, as they claim), then the bank (which you've replaced Peter with in your example), buys the bond via crediting Treas's TT&L account (inside money).

      But even if that's not the usual way the bank buys bonds, all you can claim is that in your cherry picked example, inside money doesn't come into play on the front end, however, it's still just the banks replacing Peter (but by borrowing "outside money" directly to do it). Well, as I've demonstrated, it's just as easy to cherry pick an example where inside money DOES come into play on the front end (e.g. bank as intermediary or Treasury Direct).

      Delete
    11. ... and even if the banks directly borrow outside money to make a bond purchase... guess how they repay it? One common method is buy attracting customer transfer deposits... i.e. inside money (paired w/ reserves). So even in that case it's not cut and dried... you could easily interpret that as the bank funding its bond purchase later w/ inside money!

      Delete
    12. phil, sorry to beat this to death, but let's imagine for a moment a world w/ no reserve requirements and no cash (for simplicity), and where the Treasury, after acquiring funds, spends them all immediately. Say we start with clean balance sheets for all players. Now say $T bonds have been auctioned by Treasury, of which $F have been purchased by the Fed on the secondary market, and $B have been purchased by the banks, the remaining $N = $(T - F - B) being owned by non-banks. Also assume that $L in loans have been made to the non-banks by the banks. Further assume that no bonds have yet reached maturity and that no private loans have been re-paid. Use whatever low level ordering of events that you like, I claim the resulting balance sheets will look like this:

      Tsy: A: 0, L: $T, NE: $T
      Fed: A: $F bonds, L: $F reserves, E: 0
      banks: A: $(L+B+F) L: $(L+B+F), E: 0
      non-banks: A: $(L+B+F) deposits + $N bonds, L: $L, E: $T

      Agreed?

      So now given my assumption that non-banks hold all the private sector T-bonds (B = 0) we have:

      non-bank deposits = $(L+F)

      Which matches my statement "Only to the extent..."

      If I don't make that assumption, I have to modify my statement like this:

      "Only to the extent that the Fed & banks hold Treasury bonds as assets can you say that private bank deposits in the private sector are backed by "loans to the government" (Treasury bonds) rather than "loans to private entities."

      Delete
  10. “Nice try though”.

    I’m not trying to pull the wool over your eyes; I’m trying to explain something using simple examples. I apologise if they’re not clear enough.

    We both know that in Pragcap-world “taking from Peter to pay Paul” means government “taking Peter’s bank deposit and giving it to Paul”.

    The idea being that funds originate within the private sector in the form of “inside money”, which the government then “takes” and somehow “redistributes” when it spends.

    Thus, the argument goes, the private sector “funds” government spending with “inside money” which it creates itself.

    I’m sorry, but this idea simply doesn’t make any sense at all.

    What we have is what’s called a “fiat money system”. This means that the base money, or currency, is a promise from the government. It’s a government “IOU”, not a commodity or anything else.

    In this system banks also create credit, which is a promise from banks to pay base money, or currency, on demand (either to a depositor, or else on their behalf). Credit is a bank IOU.

    It doesn’t matter whether the central bank is “independent within government” (Fed) or “wholly owned by the Treasury” (Bank of England), the above logic is the same.

    As such it should be immediately obvious that:

    (1) The government doesn’t need bank IOUs. It doesn’t get funds by receiving bank IOUs, as these bank IOUs are simply promises to pay its own IOUs on demand.

    (2) The government doesn’t get funds when it receives its own IOUs.


    Now, to get a more detailed understanding of the mechanics of the system we need to pay attention to operational details, such as the fact that the Treasury doesn’t get overdrafts from the Fed, that the Treasury maintains TTL accounts in commercial banks, that non-banks buy bonds, etc. Nonetheless, we should be careful when doing so not to lose sight of the basic logic above.

    ReplyDelete
    Replies
    1. phil, I understand your argument backwards and forwards, we've been over it many times. You always resort to talking about the "government" and it's IOUs. I've already granted you that if you consolidate the Fed and Treasury together under "government" and give them one balance sheet then you can talk about reserves being created when the gov spends and being destroyed when it taxes or collects interest. But even in that consolidated case, the private sector is in the loop (by design) when the government spends (on ANYTHING outside of certain financial assets on the secondary market). To buy a bridge the funds would go gov => private sector => gov => bridge. Can't we just cut out that "private sector" middle man? Sure, we could change things and do that, but the way it is now that private sector is in there by design. Have you ever thought about why the private sector was designed in to sit as what must appear to you to be an unnecessary middle man? Weird and stupid? Absolutely! When you look at it like that it is! It's clear that things could be streamlined if we just cut out that unnecessary middle man!

      Which is why Treasury and the Fed are set up with separate balance sheets and why the Fed is independent and which is why it's not as useful to see them as the same thing! When you do, you're missing the reason for why the system was designed the way it is! In reality, when the TGA is empty, Treasury turns to the private sector. It doesn't self finance, it doesn't create reserves on its own, it doesn't overdraw its account... it turns instead to the private sector for funding, either through bond auctions or taxation. Now does the private sector in turn need the Fed to facilitate funding Treasury! Yes, absolutely!

      When you see the world like that, it's quite easy to see the Peter to Paul route. I can see both descriptions clear as day, and I find less value in looking at it with your description! It just looks like a nonsensical, inefficient mess that way!

      Which is why the system was designed (w/ a separate Fed and Tsy) such the private sector would not be seen as an unnecessary middle man, but rather as the centerpiece of the whole system!

      Delete
    2. "You always resort to talking about the "government" and it's IOUs".

      I don't resort to it, it's how things are. There is a thing called the "government" and base money/ currency is a "government IOU". Instead of IOU we could say "obligation" or "liability". Same thing.

      This isn't an argument for changing the current system 'design', it's a description of the current state of affairs.

      "Have you ever thought about why the private sector was designed in to sit as what must appear to you to be an unnecessary middle man?"

      I'm not trying to say that the private sector is unimportant or unnecessary. Not in the slightest. Nor am I arguing that the current 'design' necessarily needs to be changed. It could potentially be altered, and that might be a good thing, but that's not the point of the comments I have made above.

      "the system was designed (w/ a separate Fed and Tsy) such the private sector would not be seen as an unnecessary middle man, but rather as the centerpiece of the whole system!"

      How would the Treasury taking overdrafts from the Fed suddenly turn the private sector into an "unnecessary middle man"? The US Treasury used to have an overdraft facility - was the system very different then? No. Treasuries in other countries (Canada, UK) still do have overdraft facilities - are their systems somehow very different to the US? No. The Canadian Treasury used to finance all its deficit spending by borrowing from the central bank - did this turn the Canadian private sector into an "unnecessary middle man"? No.

      Perhaps you could explain to me in more detail why you think it is important that the Treasury be stopped from taking overdrafts, minting platinum coins, printing US notes, etc?

      Delete
    3. The facts remain: before Treasury can spend it must fund the TGA, and to fund the TGA it turns to the private sector. Let me know when that changes.

      Delete
    4. "the system was designed (w/ a separate Fed and Tsy) such the private sector would not be seen as an unnecessary middle man, but rather as the centerpiece of the whole system!"

      I don't get this argument. Are you saying that if the Treasury borrows from the central bank then suddenly the private sector becomes unimportant? The only thing which makes people think the private sector as important is the fact that the Treasury doesn't take overdrafts from the Fed?

      Delete
    5. "the system was designed (w/ a separate Fed and Tsy) such the private sector would not be seen as an unnecessary middle man, but rather as the centerpiece of the whole system!"

      A few more points on this argument:

      1. The largest single holder of US debt is China - a dictatorship controlled by the Communist Party. Does this mean that the system was "designed" to make the Chinese Communist Party look like the "centerpiece of the whole system"?

      2. The only required intermediary between the Fed and the Treasury is commercial banks. So when you say "the private sector" what you really mean is "the banks".

      3. The implication of your argument is that the "private sector" is actually an "unnecessary middle man", and it only *appears* not to be because the government doesn't let the Treasury borrow directly from the Fed.

      So why does the government have to stop the Treasury borrowing from the Fed in order to make the "private sector" *look like* it's the "centerpiece of the whole system" instead of an "unnecessary middle man"?

      Delete
    6. phil, my "centerpiece" comment is not intended to be a "proof" of anything... it's my interpretation. Why do you think our system is set up so the Treasury turns to the private sector for funding? Why do YOU think that is?

      1. China... probably the private financiers who devised the system never foresaw that!

      2. banks vs private sector: yes, the banking part of the private sector must serve as the intermediary for the rest of the private sector. This is even true for Treasury Direct sales (since ultimately a bank ends up acquiring the Fed deposits which are sent to the TGA for you). Why? My opinion is because it was primarily private financiers who devised the Federal reserve system!

      3. No, to me the private sector doesn't look like an unnecessary middle man. I'm saying it MIGHT look that way to somebody that insists on consolidating the Fed with Treasury. To me it looks like the primary way the Treasury obtains funding. I think (again, my opinion, not a proof of anything!) that's what the designer of the system had in mind. If you want to think of the Treas and Fed consolidated, be my guest! I don't care. I'm not going to try and talk you out of that view.... that's a perfectly valid way to see it. I don't think it's a useful view of things.

      You and I can (almost) agree on every detail about what happens. I have the impression that we're looking at one of those optical illusions with the vase vs the two faces:

      https://encrypted-tbn1.gstatic.com/images?q=tbn:ANd9GcQnuJE9wjdrc5oin3OKfbXeYwDab3WwxmAwdfndNo9hvzrNt2F5

      and you keep insisting that it's a vase and I keep saying I prefer to see it as two faces... but for some reason it's super important for you to make me agree that it's really a vase! I don't get it!... yes, I see the vase, but I prefer the faces!! EITHER explanation results in the same borderline between black and white regions (all the low level details).

      Delete
    7. ... and another thing... you seem to think that because inside money doesn't DIRECTLY flow into the TGA to become outside money, that that makes your view the "correct one."

      An analogy in engineering is "hole flow" through a semi-conductor. What is that? Well at the lowest level, it's when electrons jump from molecule to molecule filling a space where a missing electron was. Someone might object to the idea of a "hole flow" saying (pedantically) everytime "Holes don't flow!!! It's just the electrons moving! STOP SAYING THE HOLES FLOW!!! That's NOT what happens... ONLY the electrons move! There's no such thing as holes flowing!! It can't happen!"

      Actually perhaps a better analogy is an electric signal "flowing through" a capacitor! Since the electrons don't ACTUALLY flow between the plates (instead the electric field induces a flow on either side), I have a feeling you'd be there in class jumping up and objecting every time "The electrons can't FLOW between the plates... it CAN'T HAPPEN so stop saying that there's a current through the capacitor! It can't HAPPEN!!!" Hahaha!

      The fact is "holes flowing" or "current flowing" is used because it's a USEFUL CONCEPT! It's like you can't (or won't) see the duality of the thing. I don't really care that you don't! I'm not trying to convince you to adopt my view... both your view and my view explain WHAT HAPPENS, so it's irrelevant to me.

      Try this as a mental exercise: consolidate the balance sheet of the Fed with the private banks. Why?... well... say you do it because the banks own the Fed's stock. Yeah... I know that isn't normal stock... etc., fine. This is just an exercise! Now replace on the currency "Federal Reserve Note" with "Federal Reserve Corporation Note." Does that really change anything? Does it make the system dramatically different? NO! Is it a useful view? I don't think so... but probably somebody out there does... perhaps some Fed conspiracy theorist for example! Does it still explain EVERYTHING THAT HAPPENS! Yes!

      Delete
    8. The Treasury had an explicit overdraft facility at the Fed until 1981. In the eighties there was a big resurgence in right-wing politics and economics. I'd pin the current "design" on that, rather than on any cabal of financiers in the early 20th century.

      "I have the impression that we're looking at one of those optical illusions with the vase vs the two faces"

      That's the way I see a lot of economic arguments. For example: which comes first, investment or saving, demand or supply, loans or deposits, etc.

      Delete
    9. "No, to me the private sector doesn't look like an unnecessary middle man"

      Well on the one hand you acknowledge it's not actually necessary to have an intermediary between the Treasury and the Fed. The Treasury could borrow directly from the Fed, for example.

      So what is the point of the current 'design'?

      I really want to know what you think.

      What *logical* reason is there for the current 'design'?

      Delete
  11. phil, BTW, I found this:

    http://useconomy.about.com/od/monetarypolicy/f/Who-Owns-US-National-Debt.htm

    and this:

    http://www.bloomberg.com/news/2012-08-20/banks-use-1-77-trillion-to-double-treasury-purchases.html

    which don't seem to agree. But it occurs to me that I should expand "banks" into "Fed reserve holders" in that formula. The bonds they hold represents bonds NOT held by us lowly bank deposit holders (mutual funds, individuals, etc), and thus represents more bank deposits in the system (assuming TGA = $0).

    ReplyDelete
    Replies
    1. "prior to the financial crisis that followed the collapse
      of Lehman Brothers on September 15, 2008, the TGA
      mostly fluctuated in a narrow band around $5 billion while TT&L balances varied more widely. In contrast, since the fall of 2008, TT&L balances have fluctuated in a narrow band around $2 billion and the TGA has varied widely."

      http://www.newyorkfed.org/research/current_issues/ci18-3.pdf

      Delete
    2. "Only to the extent that the Fed & banks hold Treasury bonds as assets can you say that private bank deposits in the private sector are backed by "loans to the government" (Treasury bonds) rather than "loans to private entities."

      When the Treasury spends, this creates deposits in the banking system. Those deposits are "backed by loans to the government".

      There are different ways in which these deposits could end up being "deleted" from the system in aggregate, but we can't really say how many "remain in existence" and how many "get deleted".

      Unless you know that, then you can't really say anything about how many deposits in the system are "backed by loans to the government", or not, with any degree of accuracy.


      Delete
  12. phil, all I did there was write out in words the formula YOU came up with + the component from private loans. I thought you'd be happy! Here's where you came up with it:

    http://brown-blog-5.blogspot.com/2013/03/list-of-ways-reserves-leave-banking.html?showComment=1367422068745#c6088224797655956344

    Here's how you stated it in that comment:

    "Non-bank deposits = bonds held by banks + reserves"

    We can alter the wording slightly and use "Fed held Tsy debt" rather than "reserves" since it's clear how they are related, and we get:

    Non-bank deposits = Tsy debt held by banks + Tsy debt held by the Fed + private loans

    You can go through my example (based on your example!) and see exactly where I got:

    non-bank deposits = $(L+B+F)

    Do you have a problem with this formula?

    Can this be made more general by expanding outside the confines of your example? Sure... but I think your example was reasonable enough for a first cut.

    ReplyDelete
    Replies
    1. "Non-bank deposits = bonds held by banks + reserves"

      This is true in that particular example. I wasn't sure at the time whether it might be correct in general, and upon reflection I think it's clear that it isn't.

      Say you have a bank deposit and you buy government bonds through Treasury Direct. When you make the payment your bank just deletes your deposit. The Fed then debits your bank's reserve account and credits the TGA. Your deposit doesn't go anywhere, it just gets deleted from the liability side of your bank's balance sheet.

      When the Tsy spends, a bank deposit is created in the banking system somewhere. That deposit was created as the result of govt spending, or is "backed by loans to the government". That deposit might get deleted somehow - for example someone might withdraw it as cash, or it might be used to repay a bank loan, or it might be used to buy a govt bond, etc. But we can't know what happens to it.



      Delete
    2. WHat I'm saying is that bank loans create deposits, but then these deposits can get 'deleted' in different ways - repaying loans, buying government bonds, withdrawal as cash etc. THe same goes for government spending - this creates deposits, which can then get deleted in different ways. So we can't really know which of the deposits in the system are "backed by loans to the non-govt" or "backed by loans to the govt".

      Delete
  13. phil, you write:

    "So we can't really know which of the deposits in the system are "backed by loans to the

    non-govt" or "backed by loans to the govt"."
    Let's not try to sort out "which of the deposits in the system" are backed by either. We

    can't say much about any particular deposit most of the time, but we can talk about the

    deposits in aggregate. As an analogy, if person x pays me a $1 and person y pays me a $1

    so now my deposit has $2 (total), I can say $1 came from x and $1 came from y. Now say I

    pay person z $1 so their deposit has $1 total. Is person z's dollar from x or from y?

    That's a nonsensical question. However, can I say that of the $2 in the aggregate deposit

    between me and z that $1 came from x and $1 came from y? Absolutely! If you don't like

    the phrase "backed by" substitute instead "corresponds to" .. I don't care! I'm just

    describing the formula.

    Now let's see if the formula holds in your updated word example.

    "That deposit might get deleted somehow - for example someone might withdraw it as cash,

    or it might be used to repay a bank loan, or it might be used to buy a govt bond,"

    I explicitly didn't cover "cash" before, but that's easy... it's simply a withdrawn

    deposit, so I can account for it by updating "non-bank deposits" to be "non-bank deposits

    & cash." Also, like I say, I prefer to state the formula as:

    Non-bank deposits = bonds held by banks + bonds held by the Fed

    Assuming here, as you did in your 1st example, that there are no private bank loans. In

    your example (which inspired me) it was clear the reserves came from the Fed's bond

    purchase and thus were equal. I prefer to keep them both bonds then as that is in line

    with my original statement (i.e. gov loans). Plus I don't have to worry about loans of

    reserves that way.

    Also, I see that you've brought up bank loans now!... thus we'll need to add in my "+

    private bank loans" to the formula:

    Non-bank deposits & cash = bonds held by banks + bonds held by the Fed + private loans

    So now let's take the example:

    Stage 1: Start:

    Tsy: A: 0, L: 0, E: 0
    Fed: A: 0, L: 0, E: 0
    bank: A: 0, L: 0, E: 0
    you: A: 0, L: 0, E: 0

    Stage 2: "Say you have a bank deposit and you buy government bonds through Treasury

    Direct."

    Tsy: A: X reserves, L: X bond, E: 0
    Fed: A: X loan to bank, L: X reserves, E: 0
    bank: A: X loan to you, L: X borrowed from Fed, E: 0
    non-banks: A: X bond, L: X borrowed from bank, E: 0

    Stage 3: "When the Tsy spends, a bank deposit is created in the banking system

    somewhere." (consistend w/ my assumption that TGA returns to 0). I'm also letting the

    bank repay the Fed here, but since I changed your "reserves" to "Fed owned bonds" this

    detail doesn't matter (try it for yourself!):

    Tsy: A: 0, L: X bond, NE: X
    Fed: A: 0, L: 0, E: 0
    bank: A: X loan to you, L: X deposit for me, E: 0
    you: A: X bond, L: X borrowed from bank, E: 0
    me: A: X deposit at bank, L: 0, E: X

    So now let's check the formula.

    non-bank deposits & cash = X
    bank held bonds + Fed held bonds + private loans = X

    Check!

    (Example continued in next comment)

    ReplyDelete
  14. (Continued from previous comment)

    "That deposit might get deleted somehow - for example someone might withdraw it as cash,"

    Stage 4a:

    Tsy: A: 0, L: X bond, NE: X
    Fed: A: X loan to bank, L: X cash, E: 0
    bank: A: X loan to you, L: X borrowed from Fed, E: 0
    you: A: X bond, L: X borrowed from bank, E: 0
    me: A: X cash, L: 0, E: X

    non-bank deposits & cash = X
    bank held bonds + Fed held bonds + private loans = X

    Check!

    "or it might be used to repay a bank loan," ... say I'm a nice guy and give you back the

    $X so you can repay your loan:

    Stage 4b:

    Tsy: A: 0, L: X bond, NE: X
    Fed: A: 0, L: 0, E: 0
    bank: A: 0, L: 0, E: 0
    you: A: X bond, L: 0, E: X
    me: A: 0, L: 0, E: 0

    non-bank deposits & cash = 0
    bank held bonds + Fed held bonds + private loans = 0

    Check!

    "or it might be used to buy a govt bond," Say I buy the bond from you:

    Stage 4c:

    Tsy: A: 0, L: X bond, NE: X
    Fed: A: 0, L: 0, E: 0
    bank: A: X loan to you, L: X deposit for you, E: 0
    you: A: X deposit, L: X borrowed from bank, E: 0
    me: A: X bond, L: 0, E: X

    non-bank deposits & cash = X
    bank held bonds + Fed held bonds + private loans = X

    Check!

    Now what if you repay your loan? Then we're back to Stage 4b (in terms of the formula).
    Or I could buy another government bond from Tsy, and Tsy then spends the proceeds
    (consistent w/ my TGA returns to 0 assumption). Say they spend on someone else:

    Stage 4d:

    Tsy: A: 0, L: X bond, NE: 2X
    Fed: A: 0, L: 0, E: 0
    bank: A: X loan to you, L: X deposit for someone else, E: 0
    you: A: X bond, L: X borrowed from bank, E: 0
    me: A: X bond, L: 0, E: X
    someone else: A X deposit, L: 0, E: X

    non-bank deposits & cash = X
    bank held bonds + Fed held bonds + private loans = X

    Check!

    So how is it "clear that it isn't" true?

    If you're talking about saying where each particular deposit came from, that doesn't make
    sense. That's why we've been using aggregates. If you're talking about being withdrawn in cash, sure we didn't cover that before, but now it's covered. If you're talking about

    repaying private loans, then we have to fold in private loans (i.e. use my formula), if
    you're talking about purchasing more bonds, then that doesn't violate the formula.




    ReplyDelete
  15. just to clarify, I'm talking about your prior statement that "only to the extent that the Fed holds Treasury bonds as assets can you say that private bank deposits in the private sector are backed by "loans to the government" (Treasury bonds) rather than "loans to private entities."

    Pretend for a moment there's no money in existence (!) and then let's say the following:

    1. Tsy sells bonds to banks, and spends. This creates a bank deposit for Mr. A.

    The quantity of bank deposits "backed by loans to the government" now equals the quantity of bonds held by banks.

    2. Mr. B takes a loan from a bank. This creates a new deposit for Mr. B. He then spends it, buying the bonds from the banks. This results in Mr. B's deposit being "deleted".

    Now the quantity of bank deposits "backed by loans to the government" equals the quantity of bonds held by Mr. B.

    ReplyDelete
    Replies
    1. phil, why go back to

      "only to the extent that the Fed holds Treasury bonds as assets can you say that private bank deposits in the private sector are backed by "loans to the government" (Treasury bonds) rather than "loans to private entities."

      I already granted you that it's more accurate to write:

      "only to the extent that the Fed & banks holds Treasury bonds as assets can you say that private non-bank held bank deposits & cash in the private sector are backed by "loans to the government" (Treasury bonds) rather than "loans to private entities."

      "& cash" takes into account the cash (which you'd brought up) and the "& banks" makes it more generally true (w/o "& banks" it's just an approximation, valid only when banks hold a small amount of T-bonds).

      Other than that, the only thing I can see that must be bothering you is the phrase "backed by."

      Assuming Tsy spends all the TGA (i.e. TGA = 0), then at each step in your example my formula still holds:

      private non-bank held deposits & cash = private bank loans + Fed & bank held Tsy debt.

      Do you agree? Starting from all balance sheets being empty, can you find a counter example? Your item 2. above was not a counter example, because the bank loan to Mr. B still exists, so the formula still holds. What you wrote is not generally true.

      We could write this out in words w/o "backed by" if that would make you happier. We could say:

      "Private non-bank deposits & cash in excess of private loans is equal to the Treasury debt held by the Fed and the banks."

      Personally I don't feel that offers as much insight as the "backed by" sentence, but whatever floats your boat! We could also expand to get at the same idea w/o "backed by":

      "Assuming all Treasury funds have been spent back into the private sector (i.e. the TGA is empty) aggregate private non-bank deposits & cash can be divided into two components, corresponding, respectively, to two components of debt. The first component of debt is private bank loans (obviously). The second debt component is Treasury debt held by the Fed & the banks. This is because (unlike non-banks) the Fed and banks do not purchase Treasury debt from existing commercial bank deposits. Thus private non-bank deposits & cash in excess of private bank loans must equal the Treasury debt held by the Fed and the banks."

      Delete
  16. That would be a kind of accounting identity, but it doesn't necessarily tell us much about the line of causation.

    For example someone might say: "bank loans equal deposits, therefore only to the extent that customers make prior deposits can there be subsequent bank loans"

    This is the old fallacy that banks sit around waiting for customers to make deposits before they "loan them out".

    Or, for example: "saving equals investment, therefore only to the extent that people have previous savings can there be subsequent investment"

    This is the same sort of fallacy, the idea being that people have to save more now so there can be more funds available for investment in the future.

    Both of these fallacies take a basic accounting identity and misinterpret them to reach an incorrect conclusion about causation.

    In my example above, in step (1) government spending results in the creation of a new bank deposit. This is a deposit "backed by loans to the government".

    In step (2) another bank deposit is created by a private bank loan to Mr. B. This deposit is then used to buy govt bonds.

    After steps (1) and (2) the deposit "backed by loans to the government" still exists, but the deposit "backed by loans to private entities" doesn't. So the quantity of deposits backed by loans to the government now equals the quantity of govt bonds held by Mr. B.

    However the quantity of deposits backed by loans to the govt also equals the quantity of “loans to private entities.”

    As such, you want to argue that the quantity of deposits “backed by loans to the government” equals zero, because the quantity of deposits equals the quantity of “loans to private entities”.

    It should be clear from the example, however, that such an interpretation is incorrect. Basically what you’re doing is taking an accounting identity and misinterpreting it, to reach an incorrect conclusion about causation.


    JKH talks about some other ways in which deposits might end up being ‘deleted’ in this post: http://monetaryrealism.com/loans-create-deposits-in-context/ These include buying bank stock and bank debt.

    ReplyDelete
    Replies
    1. I'll take that as a "yes" then. The words "backed by" are what bothered you! I'll also take it as a "no" as to a counter example. You don't challenge the equation... uh, oops, "accounting identity" then? Call it what you will, it wasn't obvious to me when I first sat down to think about it... nor you either apparently, based on your comments about it a month ago and its "murky" nature.

      I never intended to trace the causation of every deposit in the system. That quickly becomes an impossible task, akin to asking if "person z's" dollar came from x or y in my previous example! I wanted to instead be able to say something about those bank deposits that were out there, in aggregate, in relation to debt.

      Here's another variation:

      "Only to the extent that the Fed & banks holds Treasury bonds as assets are there private non-bank held bank deposits & cash in the private sector in excess of "loans to private entities.""

      To me it's quite obvious why there is a component of bank deposits equal to the quantity of private loans. As long as TGA = 0, then those deposits don't go away (in an aggregate sense) unless the loans are repaid!

      What I wanted to do was describe the remaining component of private bank deposits... and that's the part that's equal to the Fed & bank held Treasury debt. Why is this 2nd component equal to that and not the full amount of Treasury debt? That's what I was getting at!

      I don't think what I wrote gives a devastating fallacious impression of how the system works!

      Delete
    2. In my example, after step (2) the quantity of deposits "backed by loans to the government" equals the quantity of bonds held by Mr. B, whilst the quantity of deposits "backed by loans to private entities" equals zero.

      Do you disagree?

      Delete
    3. "your comments about it a month ago"

      Yeah I apologise for my slow responses and the extended nature of this debate. Things get in the way.

      Delete
    4. phil, thanks for the debate. I won't be available myself for a while starting today! (I've got to lay off the blogging for a bit).

      Regarding your example, I can see why you say what you do, so I won't disagree w/ you.

      Take care!

      Delete
    5. ... I changed the post accordingly. In an aggregate sense we can say that when the Fed & banks (aggregated together for this purpose) buy Tsy debt (from Tsy or non-banks) they create bank deposits (under the assumption of deficit spending: i.e. the TGA is empty). When non-banks buy Tsy debt from the Fed & banks, non-bank bank deposits decrease, but when they buy from Tsy, there's no net change in bank deposits (again under the assumption of TGA = 0).

      Delete
  17. Tom,

    I really recommend this paper by Scott Fullwiler:

    "Interest rates and fiscal sustainability" (2006)

    http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1722986

    It covers a lot of the topics we've discussed here, and relates them to the broader ongoing economic debate.

    I hope you can take the time to read it. Let me know what you think!

    ReplyDelete
  18. Hi Tom,

    I'm a college student who just finished my intermediate macro course and am looking to learn more about our monetary system in greater detail. This post was very helpful. Thank you very much.

    Do you mind if I ask you a question? It may seem like a stupid one, but I feel like I am missing something.

    "Now these deposits can be traded by the non-banks for the Treasuries and back again.... in any amount up to $X. Thus bank deposits can vary between $0 and $X under this trade."

    My question is, if there is $X in bank deposits, then when non-banks buy Treasuries, which subsequently are spent by the Treasury, wouldn't the bank deposit end up being $X in total after wards as well? You state later on that aggregate non-bank bank deposits do not change, which makes complete sense to me. So what do you mean by "the deposits can vary between 0 and X under this trade"?

    ReplyDelete
    Replies
    1. Hi Unknown,

      Thanks for stopping by. Appreciate the comments. Now to your question:

      Those Treasuries I'm referring to are not new Treasuries auctioned by the Treasury, but rather the $X in existing Treasuries that the banks and the Fed have already acquired. That's what I meant to convey in the preceding sentences:

      "To see why consider the following example: assume we start off with everyone's balance sheets clear (zeros for assets and liabilities), and then the Fed & banks acquire $X in Treasuries. This results in $X in bank deposits when Treasury spends."

      Sorry that wasn't more clear! Does that help? I'm trying (in words) to build up a story by imposing a certain order on events but in reality those events don't necessarily have to follow that order. My ordering is (up to the sentence you asked about):

      1. Banks buy $X of Treasury debt from the Treasury (Treasury then stops auctioning debt).

      2. Fed buys some of that debt from the banks, but the total is still $X (I didn't explicitly state this step).

      3. Treasury spends all the $X in proceeds on non-banks thus creating $X in bank deposits.

      Now at this point the non-banks can trade anywhere between $0 and $X of their deposits buying Treasury debt from the banks and/or the Fed (whoever holds them... and assuming they're willing to sell). Does that make sense?

      Later on in the narrative I introduce more steps:

      4. Non-banks take loans from banks

      5. Treasury now auctions more Tsy debt again, but this time only to non-banks.

      Like I said, I selected this ordering just to explain the formula, but you could have selected another. Even here the ordering of points 2. and 3. don't affect the narrative as it stands.

      Delete
    2. Unknown, I'm very glad you found this blog helpful, but just as a warning: I'm NOT an expert on this stuff. Everything I've learned has been from people like Cullen Roche at pragcap.com, JKH at monetaryrealism.com and "Joe in Accounting" and even "phil" who comments in both places (phil and I have a lengthy debate above, as you may have noticed, during the course of which I learned a lot). I'm interested in this stuff so I'm making an attempt to learn it on my own, but I have no formal education or professional experience in it!.... so don't bet your grades on things you've "learned" on this site! I may well be correct, but I'd try to double check if it's important (and be sure to come back and let me know where I've gone wrong if you find out the truth!). See my "mission statement" tab for more info on my motivation.

      Regarding that formula in particular, that's something I worked out for myself because I was literally trying to relate the concept of "loans create deposits" to the deposits and "loans" that exist in the system. Of course the system as presented here is pretty simplified: no GSEs, foreign central banks, foreign trade, etc.

      OK, thanks again!

      Delete
    3. Unknown, shoot I forgot step 6. above:

      6. Treasury spends proceeds of Tsy debt sale to non-banks on non-banks, thus creating more deposits for non-banks.

      So you're absolutely correct, I'm always assuming Tsy spends all the proceeds thus putting deposits into the hands of the non-banks, but at the step you asked about there's no proceeds since non-banks are buying existing debt.

      Delete
    4. Shoot, still didn't quite enumerate it all, but I think you get the picture. I think I can simplify that paragraph a bit by eliminating the Z variable and associated steps.

      Delete
  19. Hi Tom,

    Thanks so much for that clarification. That makes sense to me. Though you claim to not be an expert, you are still very knowledgeable. Thanks again for sharing. You can be sure I will be back with more questions in the future!

    I do follow Cullen Roche's blog -- I find MR to be very interesting.

    It was funny to me how you had to address me as Unknown. Oops - should have provided my name! It is Anne.

    ReplyDelete
    Replies
    1. Well then, Hi Anne. Best of luck with your classes!

      Delete
  20. "...banks don't "lend out" those excess reserves to the public. That's a common misconception people have and is what really inspired this post."

    So...in what way do these QE-driven excess reserves stimulate the economy and/or expand money supply. As I see it, when the bank clears a bond trade to the Fed, it is paid in [excess] reserves. It pays the bond seller by crediting a deposit account. No new money. What happens (or is supposed to happen) next?

    ReplyDelete
    Replies
    1. The do expand the money supply because reserves are money. However, in one very important respect they can violate a core property of money: guaranteed 0% nominal rate of return. Why? Because the Fed decided to pay interest on them (IOR).

      The idea is to drive down the demand for money. It's not at all clear that this works. Some people say that IOR (positive IOR) is the reason it didn't work well.

      Personally, I don't really know.

      In the extreme case it's hard to see this general mechanism not working (i.e. the Fed buying up the whole world). But apparently they can add LOTS of base money (those deposit accounts for non-banks held at commercial banks don't qualify technically as "base money"), and not much happens.

      Delete
  21. I would like to say that this blog really convinced me, you give me best information! Thanks, very good post.
    buy perfect money with credit card
    Keep Posting:)

    ReplyDelete