Monday, March 11, 2013

Banking Example #4: Quantitative Easing

See Example #4.1 for a simplified comparison of the two kinds of QE. The example below focuses on the second kind, which is more prevalent.

Setup: one central bank (CB), one commercial bank (A), and one person x. Reserve requirements are 10% of deposits, but there are no capital requirements. Person x has previously obtained a $100 loan from Bank A*, and Bank A obtained $10 in required reserves by borrowing it from the CB. Person x also has a $200 Treasury bond, which he intends to sell** (in this case to the CB during a CB Quantitative Easing (QE) operation):

Initial balance sheets :

CB
Assets Liabilities
$10 loan to A $10 reserve deposit for A

Bank A
Assets Liabilities
$100 loan to x $100 deposit for x
$10 reserves (required) $10 borrowing from CB
Negative Equity Equity
----------------------- $0

Person x
Assets Liabilities
$100 deposit at A $100 borrowings from A
$200 Treasury bond -------------------------
Negative Equity Equity
----------------------- $200


Balance sheets after CB purchases $200 Treasury bond from Person x during QE (and in the process credits Person x's bank, Bank A, with $200 of reserves), and Bank A repays the $10 of reserves it borrowed from the CB:

CB
Assets Liabilities
$200 Treasury bond $200 reserve deposit for A

Bank A
Assets Liabilities
$100 loan to x $300 deposit for x
$200 reserves ($30 required, $170 excess) -------------------
Negative Equity Equity
------------------------------------------- $0

Person x
Assets Liabilities
$300 deposit at A $100 borrowings from A
Negative Equity Equity
----------------------- $200


Take note that QE did not change the equity on anybody's balance sheet, including the CB's. Bank A ends up with $170 of excess reserves (up from $0 initially), but its equity (assets - liabilities) remains at $0. Also keep in mind that I'm not showing the Treasury Department's balance sheet, on which the $200 bond would appear as a liability (I could also show an offsetting $200 in the Treasury Dept.'s Fed deposit account as an asset, supposing that money had not yet been spent). The CB and the Treasury Dept. are different entities.

Another thing to note is that although person x's equity position doesn't change immediately before or after this bond sale, his position may have changed over the course of a QE operation, since the Fed tends to buy enough of the bonds in the market so as to affect the price of Treasury bonds (and some would argue, all financial assets in general). However even the existence of this asset price effect is somewhat controversial. Of course Treasury continues to auction new bonds. The overall effect is beyond the scope of this post, but others have covered it, such as pragcap and FTAlphaVille.

*Obviously this initial loan is not required for the example. I included it to parallel a ZeroHedge article with a chart demonstrating how loans had previously (before QE) tracked deposits, but since the start of QE deposits have grown faster than loans. As Cullen Roche has pointed out on pragcap, the reason for this divergence is not as mysterious as the ZH article implies it is (it's basically for the reason illustrated in this example). Cullen has also stated that about 20% of Treasury debt is held by commercial banks.

**Note also that the bank could also sell Treasury bonds directly to the Fed if it owned any. In that case the bank's balance sheet would not expand. However, I've chosen to show what I believe is the more typical scenario, wherein the Fed purchases the bonds from a private non-bank entity (perhaps using the bank as an intermediary for the sale).

Cullen Roche has cited a NY Fed SOMA desk official he spoke with to support this assertion about who the primary bond investors are (in this case the quote actually refers to who the ultimate buyer is at the Treasury auction):

“The primary way dealers finance their bond purchases is in the repo market. So here is one scenario. Funds are wired from the dealer’s account at its clearing bank to Treasury on issuance day. During the day, the clearing bank provides intraday credit to the dealer, so the dealer is borrowing from the bank. That same day, the dealer enters into a repo, pledging the newly acquired Treasury as collateral. The other side of the repo is likely to be a money market mutual fund or other money market investor. Therefore, by the end of the day, and for the overnight period, the money market investor is effectively funding the dealer’s position. Of course, there are a variety of ways in which positions can be funded, but the repo market is the key one.”

23 comments:

  1. Tom, what do you know about the Repo mkts?

    ReplyDelete
  2. After, QE, the bank now has excess reserves that it can use to speculate on equities thanks to Uncle Ben. Booyah! ;-)

    ReplyDelete
    Replies
    1. It can certainly have an expanded balance sheet (as I demonstrate here), but clearly that doesn't necessarily mean more equity. The bank must still meet capital requirements, and in general maintain an acceptable CAMELS score.

      Delete
  3. Federal Reserve Act: A license to counterfeit money.
    Treasury borrows money from Federal Reserve and it goes like this.
    Treasury issues $100 T-bond to the Fed who orders $100 Federal Reserve Note to be printed by Treasury, cost to Fed: 2 cents.
    I need a racket like that one.

    ReplyDelete
    Replies
    1. "Treasury borrows money from Federal Reserve and it goes like this.
      Treasury issues $100 T-bond to the Fed who orders $100 Federal Reserve Note to be printed by Treasury, cost to Fed: 2 cents. I need a racket like that one."

      Actually the Fed is not allowed to buy Tsy debt from Tsy.

      All our money is based on debt, thus there's always liabilities and assets created in the money creation process. That how fiat money (the world's current money system) works. It's all IOUs. It's not an irreducible pile of gold, coins, or paper notes somewhere: one entity's debt is another one's money. To have money requires than debt exist. This is the oldest form of money, with records of credits and debits going back to ancient Sumer more than 5000 years ago on clay tablets (see David Graeber)... ..LONG before coins were even invented. Barter was present back then too, but it was never the main event. Day to day transactions in the ancient world and in modern primitive societies are debt oriented: not barter oriented. Coin oriented monetary systems have come and gone, so they have historically been important in several eras... but the most recent era came to a close on 1971. I think we're in for another long haul with debt based money.

      So actually you DO have that racket!! You can write as many IOUs as you want... the trick is getting other people to accept them.

      Delete
  4. Great Example,

    Would you mind taking it to the next step, when Trs is issuing additional 200usd, which person X requires, and sells directly to CB. My balance sheets dont add up!

    Because Person X needs to buy 200 bonds, ie decrease deposits by 200, which decrease reserves by 200. And the sells bonds to Trs, add 200 to deposits, which is adding 200 to reserves, ie status que pre transaction... BUT balance sheet of CB grows????

    ReplyDelete
    Replies
    1. I'm not sure I understand what you are requesting. Is it this?

      1. x buys $200 of bonds from Tsy
      2. x sells $200 of bonds to CB

      After step 1., balance sheets are:

      Tsy (just the change):
      Assets: +200 in Fed deposits
      Liabilities: +200 T-bond
      Equity: +0

      CB:
      Assets: 200 T-bond, 10 loan to A
      Liabilities: 200 deposit for Tsy, 10 deposit for A
      Equity: 0

      Bank A:
      Assets: 10 reserves, 100 loan to x
      Liabilities: 10 borrowing from Fed, 100 deposit for x
      Equity: 0

      Person x:
      Assets: 100 deposit at A, 200 T-bond
      Liabilities: 100 borrowing from A
      Equity: 200

      After step 2.:

      Tsy: (no change)

      CB:
      Assets: 400 T-bonds, 10 loan to A
      Liabilities: 200 deposit for Tsy, 210 deposit for A
      Equity: 0

      Bank A:
      Assets: 210 reserves (30 req + 170 ex), 100 loan to x
      Liabilities: 10 borrowing from Fed, 300 deposit for x
      Equity: 0

      Person x:
      Assets: 300 deposit at A
      Liabilities: 100 borrowing from A
      Equity: 200

      Yes, the CB balance sheet grows. Now let's suppose that all loans are paid back (to Fed and to A):

      Tsy: (no change)

      CB:
      Assets: 400 T-bonds
      Liabilities: 200 deposit for Tsy, 200 deposit for A
      Equity: 0

      Bank A:
      Assets: 200 reserves (20 required + 180 excess)
      Liabilities: 200 deposit for x
      Equity: 0

      Person x:
      Assets: 200 deposit at A
      Liabilities: 0
      Equity: 200

      Now assume that Tsy spends the 200 on person x:

      Tsy (just the change):
      Assets: -200
      Liabilities: +0
      Equity: -200

      CB:
      Assets: 400 T-bonds
      Liabilities: 400 deposit for A
      Equity: 0

      Bank A:
      Assets: 400 reserves (40 required + 360 excess)
      Liabilities: 400 deposit for x
      Equity: 0

      Person x:
      Assets: 400 deposit at A
      Liabilities: 0
      Equity: 400

      Don't confuse the Fed with Tsy: they are different! Check out Example 4.1:

      http://brown-blog-5.blogspot.com/2013/08/banking-example-41-quantitative-easing.html

      Delete
    2. There's an interactive set of balance sheets here (about midway down):

      http://brown-blog-5.blogspot.com/2013/08/banking-example-11-all-possible-balance.html

      QE is increasing F (Fed purchases T-bonds from public, like in this example)

      QE where Fed purchases T-bonds from bank is different: in that case increase F by $X and decrease B by $X

      Assumption is that Tsy spends every $ on public.

      Delete
  5. Thanks you so much, very appreciated!

    Some follow up´s

    So the Private sector´s wealth is increasing by budget deficit (assuming wealth is Person X Equity, and assuming Trs bonds is budget deficit)? But this wealth also increases the liability side of CB?

    In the next step, let´s call it FED EXIT, FED desire to take down the 400 in deposit to A to zero, by getting rid of 400 bonds. How would that effect the Balance sheets above?

    Im thinking about getting rid of bonds in two ways;

    1, Bonds mature, and Trs pays FED directly 400 by issuing New Bonds Worth 400, being sold to Person A

    2, Selling in bondmarket, directly to Person A. Trs is not involved.

    ReplyDelete
    Replies
    1. "So the Private sector´s wealth is increasing by budget deficit (assuming wealth is Person X Equity, and assuming Trs bonds is budget deficit)?"

      Yes. Take a look at Ex 11 again:

      http://brown-blog-5.blogspot.com/2013/08/banking-example-11-all-possible-balance.html

      For simplicity make the following assumptions:

      Ut = 0 = No unspent funds at Tsy
      D = 0 = Bank equity distributed to public (shareholders)

      That's the situation in the colored interactive spreadsheet in the middle of the post. I also make those assumptions in the uncolored table called "Public (simplified)" towards the bottom. In "Public (more simplified)" I also assume:

      C = 0 = no cash in circulation

      You can see that the public's equity = T = Tsy's negative equity = national debt (accumulated budget deficit) = Tsy bonds sold by Tsy to CB, banks, and public (excluding Tsy bonds sold to foreign CBs and to non-Tsy gov agencies (like Social Security) and ignoring mortgage debt owned by CB). I look at mortgage debt in #11.1 and non-Tsy gov in #11.2. Next up: foreign sector.

      "But this wealth also increases the liability side of CB?"

      Not necessarily. Play w/ the interactive spreadsheet or look at the expressions: Set Ut = D = C = F = B = 0. You'll still see that the public's equity = T = Tsy's negative equity. F = 0 means CB hold no Tsy debt. Now again, from Ex #11 (simplified):

      public's equity = T = public's money stock + public's Tsy debt holdings - public's borrowings from banks = T

      But public's money stock = public's bank deposits + public's cash = L + B + F = bank loans + Fed held Tsy debt + bank held Tsy debt.

      I show how this can be in Example #8 and Example #6.

      "In the next step, let´s call it FED EXIT, FED desire to take down the 400 in deposit to A to zero, by getting rid of 400 bonds. How would that effect the Balance sheets above?"

      Tsy: unchanged

      CB: clear (all entries $0)

      Bank: clear

      Person x:
      Assets: $400 Tsy debt
      Liabilities: $0
      Equity: $0

      I illustrated your 2. above. 1. is different because that's more Tsy deficit spending, which ultimately adds to the equity of Person x (again, I'm just reasoning from the formulas in Ex 11). I'm taking it one step further though: you're leaving it at Ug = 400. I'm assuming Ug = 0 (Tsy spends everything), which gives Person x another $400 in equity, for a total of $800 = new total Tsy debt.

      Person x:
      Assets: $400 deposit (after Tsy spends), $400 T-bonds
      Liabilities: $0
      Equity: $800

      Now if x used it's $400 deposit to purchase the other $400 in T-bonds, it would just have $800 in T-bonds and the CB and Bank A BSs would be clear.

      Delete
    2. Be somewhat careful though of saying that the public's wealth is the Tsy's debt. In this simple way of accounting that's true, but there's more to it than that:

      http://pragcap.com/yes-government-deficits-equal-private-surpluses

      Delete
  6. Very interesting thank you

    Do you see any issues, like crowding out effects, with a Fed exit, except for higher rates, when/if they decide to clean up balance sheet, ie reducing reserves?

    ReplyDelete
    Replies
    1. I really don't have a good feel for that. I'd say "no" only because I've read articles by both Cullen Roche and Scott Sumner that say it shouldn't be a problem, and they sounded reasonable at the time (Roche & Sumner don't agree on a lot, but they agree on this).

      Some hyperinflationists think it will be a problem. They think that rates going up on a Fed "unwind" will cause the Tsy to not be able to pay the interest on newly issued debt and thus cause a positive feedback loop adding to more problems. See Vincent Cate's "How Fiat Dies" blog for one such amateur hyperinflationist's opinions. There are several arguments against this that I think are strong:

      1. Fed unwinding coincides with improving economy and improving NGDP and GDP and tax revenues... so any increases in interest payments the gov much put up with are offset by higher tax revenues (w/o raising tax rates).

      2. The Fed can control any part of or the entire yield curve if it wants. It happens to target the overnight rate for reserves, but if they wanted to lower the rate on 10 year notes they could by buying them (while still selling other debt). This means the Fed & Tsy could act in concert to allow the Tsy to sell new debt at low interest rates. Also as long as excess reserves are much greater than $0 (ER >> 0) the overnight rate will be pushed down to the IOR rate, currently set to 0.25% (regardless what the rest of the yield curve does). Thus until the Fed unwinds completely, ER > 0, and the federal funds rate (FFR) will be the IOR rate (low!)... thus the Tsy (if long term rates are too high) could simply issue short term debt.

      I don't have the links for Cullen's articles about this, but I do have Sumner's:

      http://www.themoneyillusion.com/?p=21786

      http://www.themoneyillusion.com/?p=21999

      I don't think Sumner makes the above arguments that I make, but his are pretty good too as I recall. Sumner doesn't look at this blog... because he thinks that banks are NOT important to macro (a very neo-classical view). Well, that's one reason: the other is I'm a complete amateur, so why should he pay attention to me? He has looked at Cullen's posts (which is kind of my intellectual basis for most of this stuff), and there's some things they can agree on, but an awful lot they don't. So read Sumner for an entirely different perspective.

      Delete
  7. Thank you, I will examine these links.

    I know we simplyfying a lot, and I'm definitely an amateur in this field. But I can´t get rid of the idea that if Fed would not have been running QE, the private sector would not have the option to convert bonds into cash, making it possible to buy and later sell even more bonds.

    Take your first example. Person X had 100 in deposits and 200 in bonds. Without QE, X could only have bought 100 in new bonds from Trs. The budget deficit would have been constrained by 100? (in your specific example). After Trs would have spent 100, X deposit would have grown by 100. Ending up with 300 bonds and 100 in deposits, total assets 400.

    BUT with QE, X could sell the bonds and get 200 in cash, now holding 300 cash, buying new 300 bonds, selling those and collecting 300 in cash, now holding 300 cash again. After deficit being spent, collecting 300 new cash, ending up with total 600 in cash deposits.

    With QE FED is now holding 500 in bonds (200 old ones and 300 new ones). The difference between QE and NO QE is the potential size of deficit spending, 100 vs 300. The same difference as for the asset side of X, 600 vs. 400. But also the Equtiy side 500 vs 200. On top of that it changed the composition; , X is sitting on 600 in cash deposits, instead of 100 deposits, 300 bonds. Big difference. No wonder X is looking for higher yielding opportunities in this situation.

    In practice, running a significant deficit without QE would have caused crowding out effects and pushing up rates. Ok, in practice it would never have happened either because the Fed would have stepped in.

    So the size of deficit is affecting X equity positively, no doubt. But thanks to QE it grew much faster than otherwise would have been possible? And it affected the composition of assets of the private sector.

    FED does NOT finance the deficit, ie the public debt. The private sector did. A complete FED exit, ie taking down reserves to zero, and equally downsizing bond position, would show that? X would end up with bonds, instead of cash deposits. Higher rates and crowding out effects most likely (since X is not holding cash in practice, X holds other bonds and stocks and stuff)

    Thanks to a multi-year loose fiscal policy, the private sector benefited by adding equity (by adding deposits) Thanks to QE this process was much faster than it should have been without QE.
    Hypothetically, Say Trs would like to get debt free, Trs could tax X by 500 (the same amount as outstanding bonds). X is reducing deposits by 500. Equity is reduced by 500. X being left with 100 deposits, and 100 in bankloan. Trs pays FED 500, who can shrink balance sheet to zero.

    ReplyDelete
  8. You write:

    "So the size of deficit is affecting X equity positively, no doubt. But thanks to QE it grew much faster than otherwise would have been possible?"

    Hmm, that would imply that there's a difference in the rate Tsy could sell $T bonds w/ and w/o QE. Remember the simple model in Ex 11 with Ut (unspent Tsy funds) always at $0. Then public's equity change is equal to the deficit, or $T

    There's no way around it: Tsy deficit = public's equity increase

    You might be correct, but that would imply the Tsy might have had a failed bond auction w/o QE. I have a hard time believing that.

    You are certainly correct that QE changes the composition. It's all right there in those simplified BSs: try it yourself: set Ut = D = 0 in the formulas. Then all becomes clear.

    Play w/ the spreadsheet: it automatically handles the transition from case 1 (ER > 0) back & forth to case 2 (ER = 0), but the public's BS doesn't change in either case: it's the same formulas.

    You write:
    "The difference between QE and NO QE is the potential size of deficit spending, 100 vs 300. The same difference as for the asset side of X, 600 vs. 400. But also the Equtiy side 500 vs 200."

    I'm not sure I agree. Say F = 0 (i.e. CB buys no bonds). Then x (pubilc) could use his $100 deposit 4x, to support a larger deficit. I don't even think the rate of deficit growth would have been effected, since Tsy would spend the funds quickly each time. The rate could be exactly the same (Ut's downward slope is the same each time... but it gets discontinuous bumps up each time a bond auction is held).

    Plus we're ignoring the role of bank held debt here (B). This is a small percentage, but it too contributes to the public's composition of assets in the same exact way as does the Fed:

    public's stock of money = L + B + F

    To sum up, I agree w/ your statement that QE (F) affects the composition of the public's assets and this may have other effects (e.g. portfolio rebalancing). But I don't agree that QE has allowed for a faster growth in T which solely determines the public's equity.

    It's possible QE COULD help T to grow larger. I can't prove that it didn't... but that implies that w/o it Tsy couldn't have sold all those bonds which would have been a disaster. Tsy has NEVER had a failed bond auction, so far as I know.

    ReplyDelete
    Replies
    1. ... and regarding the interest rates: certainly QE has ensured that the FFR = IOR, but the Fed could have kept FFR wherever it wanted w/o QE. Also, the empirical evidence suggests longer term rates (e.g. 10 yr Tsy) went UP during periods of QE and fell again between those times.

      Delete
    2. I strongly urge you to look at example #11: for two reasons

      1. It helps me make sure that everything there is correct :D

      2. If it is correct (which I think it is) then it covers all possible orders of operations here that we're speculating about... (plus you can plug actual numbers in the spreadsheet to see what happens)

      For any valid combination of numbers in the spreadsheet (I don't do range checking except to differentiate between cases 1 & 2... so it's up to you to make sure they are valid via the table at the top) I think you can check the results and lay out a scenario of how you could have gotten there via the typical operations: Tsy sells debt, banks by, or public buys, and then Fed buys from each of them, etc. Throw in public borrows and public repays some or all, etc.

      Delete
    3. Also, I think invalid entries will be obvious: negative numbers appearing anywhere except for Tsy's equity is the only evidence I can actually imagine.

      Delete
  9. The thing with trs deficit is to suppport growth right? Especially when private sector is weak.

    My hypotehesis was, without QE, Trs would have to pay up much more, in terms of interest, to be able to sell the same amount of bonds. The Rising rate of bonds would affect all other assets negatively... which in the end hurts growth... which was the main purpose from start to support!

    ReplyDelete
    Replies
    1. Ah... but QE actually resulted in 10 yr rates increasing, right?

      There's another issue: Fed remits interest payments to Tsy for Fed held Tsy debt, which does help overall expenses. But the evidence seems to indicate that yields moved up with QE.

      Delete
    2. Counterfactual is a hard thing, maybe rates has been even higher without QE? I would say that buying enough will make the yields as low as needed. No reason to reckon the process would not be nothing but monotonic (http://en.wikipedia.org/wiki/Monotonic_function)

      Delete
    3. I agree that a counterfactual is a hard thing. But from what I understand when QE programs ended, rates came down. When they started, rates when up and stayed up for the duration. This seems counterintuitive on one level, but then again it's hard to say: perhaps the Market Monetarist story is closer to the truth here, and the market anticipated inflation when QE programs started and disinflation or deflation when it ended. So who's to say which result is "counterintuitive?"

      Delete