Thursday, January 10, 2013
John Lounsbury — The Terrifying Danger of the Trillion Dollar Coin
Excellent post by John Lounsbury shows the real implication of TPC, the option of government self-financing and the consequent ending of the public subsidy to finance in the form of operationally unnecessary interest payments on consolidated non-government saving of net financial assets in aggregate and with it ending the so-called government intertemporal budget constraint.
Global Economic Intersection
The Terrifying Danger of the Trillion Dollar Coin: No One is Talking about the Bottom Line with the Platinum Coin
John Lounsbury | Managing Editor and Co-founder of Global Economic Intersection
Subscribe to:
Post Comments (Atom)
36 comments:
Tom the Dancing Bug has a nice one out today
http://boingboing.net/wp-content/uploads/2013/01/1120cbCOMIC-platinum-coin1.jpg
His point number three doesn't make any sense to me:
"3. The Federal Reserve increases the reserves of the primary dealers to cover their money transfer to the government."
First, not all PDs are banks, and only banks in the Federal Reserve system can have "reserves".
Second, even if the PD is a bank, their reserves would be debited to pay for the bond, not increased.
The PD's are 16 big banks. They can borrow reserves from the Fed to purchase bonds for their inventory which they then sell on. The PD's are the actual market makers rather than the Fed, which just auctions the tsys as the Treasury's fiscal agent, so the PD's inventories expand and contract with rising and falling demand. It's a system designed to maintain high liquidity for tsys.
Some of them are banks and some of them aren't.
Here's the list:
http://en.wikipedia.org/wiki/Primary_dealer#Current_list
Also ... if what he's describing are loans from the Fed directly to the PDs, he doesn't make that clear. Also pretty sure that this was possible only for a short period during the financial crisis, under the Primary Dealers Credit Facility.
Ken,
Only excess reserves can be used to buy treasury bonds. Excess reserves are a function of FED policy. If the FED has not previously left excess reserves in the banking system they will be forced to add them at the time of treasury auction. When the PD’s go to obtain needed reserves to fulfill their auction duties – regardless of whether or not they get the reserves directly from the FED or from the money markets – the FED must supply them or loss control of its target rate.
Ken, I believe that they are all investment banks or akin to what we used to call investment banks in the US before the crisis. They are large financial institutions.
Adam1
I believe that the rb are taken out of the system when the PD's sell the bonds along. The PD's can borrow directly from the Fed to finance the deal, since the collateral from the loan is the tsys they are buying. As the Fed acts as the fiscal agent for the Tsy, so too the PE's act as brokers for the tsys into the market. They are the market makers. That's the way I understand it anyway.
@Tom,
The selling of bonds in the secondary market by the PD is just an interbank clearing process. The reserves leave the “system” when the Treasury sweeps its accounts with the PD’s after the auction. Presumably the Treasury is selling bonds because it needs to add funds to its FED account and hence removes the reserves with its post auction sweep.
Now I will caveat that I’m not an expert at that level of procedure, but that’s how I understand it.
Maybe someone with deeper knowledge of the ops can enlighten us on this. My understanding is that the reserve drain occurs when the PD's sell the tsys into the market in exchange for rb. The deal between the PD's does not since it is a brokerage that the Fed facilitates by lending.
I should add that the auctions are generally well-subscribed or oversubscribed with the PD's so they already have orders pending at the time of the auction. It's not that the PD's borrow reserves from the Fed to take the auction and then have to sell it on piecemeal after that. The Fed uses the PD"s as its brokerage agency, so to speak, so it doesn't have to deal with all the buyers involved itself. The PD's take a small fee for the service, but small fees add up in volume.
Wikipedia — Primary Dealer
Ken and Tom -- They are all investment banks or subsidiaries of one. Only a few look like they might not be, but I looked those up and they all refer to themselves as banks or investment banks.
Unfortunately, I can't help with the operational details. :)
They may all refer to themselves as "banks", but I think that's a loose definition. There's a difference between a commercial bank with reserve accounts at the Fed and access to the discount window, and an "investment bank". Pretty sure the latter is just like any other entity in our economy, no matter what they call themselves.
Some "investment banks" converted themselves into "real banks" during the crisis, so they could get access to the discount window. But I still don't think all of the PDs are actual banks.
It's true that reserve balances ultimately buy the treasuries .... true in the same sense that reserve transfers ultimately clear all transactions in the economy that involve more than one bank.
So if a non-bank (see above) PD buys a gov't bond, reserves will be transferred to the Fed (i.e. they will vanish), but the PD's account (at a real bank) will also be debited, so I don't see that the Fed need take any special action on reserves in this case. Yes, total bank reserves will go down, but so will deposits.
I don't think that there is any standing Fed facility to finance these purchases, other than the temporary programs established during the crisis. I could be wrong, so someone who is an expert on this stuff could enlighten me.
From what I can see the Primary Dealer Credit Facility is the "discount window" for the PD's but they regularly use repotoo.
I went poking around at one point to see what the standing arrangements were for PD liquidity. What I found was along the lines of what Ken said. There was a Primary Dealer Credit Facility established in the wake of the crisis but it has since been closed and all its loans repaid.
The Term Securities Lending Facility and Primary Dealer Lending FAcility were created to ease the liquidity crisis by accepting a broader range of securities than tsys, Normally, the Fed only lends against tys as collateral.
Term Securities Lending Facility
In March 2008, the Federal Reserve announced the creation of the Term Securities Lending Facility (TSLF) under the authority of section 13(3) of the Federal Reserve Act. The TSLF loaned Treasury securities to primary dealers for one month against eligible collateral. For so-called "Schedule 1" auctions, the eligible collateral comprised Treasury securities, agency securities, and agency mortgage-backed securities. For "Schedule 2" auctions, the eligible collateral included schedule 1 collateral plus highly rated private securities. The program supported the liquidity of primary dealers and fostered improved conditions in financial markets more generally. Information on collateral policies and interest rates charged for lending are discussed in the collateral and rate setting and risk management sections of this website. Operational details of the TSLF are published on the Federal Reserve Bank of New York website. Effective July 1, 2009, Schedule 1 auctions were suspended. The TSLF was closed on February 1, 2010.
TSLF
This piece at FRBNY is the best I have been able to come up with so far, but it doesn't get into the nitty gritty of Fed lending to the PD's, but having accts at the Fed they can repo tsys.
CURRENT ISSUES IN ECONOMICS AND FINANCE
The Treasury Auction Process: Objectives, Structure, and Recent Adaptations

February 2005 Volume 11, Number 2
JEL classification: G18, G28, H63
Authors: Kenneth D. Garbade and Jeffrey F. Ingber
@Ken,
You need to think of reserve usage and drains more broadly. Interbank transactions are occurring all the time and that volume requires some said level of reserves. As transaction volume increases and decreases the minimum amount of reserves needed to sustain the FED’s target interest rate changes. Prior to paying interest on reserves and the current zero interest rate policy the FED would routinely perform OMO to keep its rate (as well as allowing intra-day overdrafts by member banks without penalty). Currently the system is awash in so much excess reserves it’s all a moot point.
Anyhow, unless a PD is obtaining reserves to fund an auction purchase directly from the FED, it must then be drawing reserves from the money markets which means it draining otherwise needed reserves for interbank clearing. This causes some type of overdraft in the system causing the FED to add more reserves. Even if bank deposits are declining the auction is causing an increase in demand for reserves because there is not a 1:1 relationship between deposits and reserves AND the target interest rate is set and defended based upon DEMAND of reserves by the system which is composed of interbank clearing needs as well as demand for reserves to meet reserve requirements (and purchase of cash and coin to support customer demands for said items).
Yes, in the end balance sheets reflect a swapping of deposits for treasury bonds, but that is after the FED has supported the auction. The above process will even be reversed as the Treasury will spend those reserves back into the system causing excess reserves to pile up – unless like now the FED policy reflects allow excess reserves to exist (or sets a floor on its target interest rate by paying interest on reserves).
Again, I don't have operational expposure or direct knowledge of what goes on at a Treasury auction. Scott would likely know though.
Well, the very purpose of selling a gov't security, by the Fed or by Treasury, is to drain reserves. In the case of Treasury, to drain the reserves its spending has addded. In the case of the Fed, to drain additional reserves.
But in both cases, draining reserves is the reason for doing it. So it doesn't make sense for the Fed to **add** reserves in support of this process.
The following from the NY Fed ... since it talks about PD accounts at "clearing banks", they are obviously not assuming that the PD itself is a bank:
"The Federal Reserve conducts open market operations with primary dealers—government securities dealers who have an established trading relationship with the Federal Reserve. So while the target policy rate is the uncollateralized lending rate between banks (fed funds), the Fed operates in the collateralized lending market with primary dealers (repo). This structure works because the primary dealers have accounts at clearing banks, which are depository institutions. So when the Fed sends and receives funds from the dealer's account at its clearing bank, this action adds or drains reserves to the banking system."
URL for the above:
http://www.newyorkfed.org/aboutthefed/fedpoint/fed32.html
@Ken, the purpose bond sales for the FED is to drain reserves. The Treasury will do it to support FED policy, but the Treasury is also not allowed to overdraft with the FED and therefore must maintain a positive balance with the FED prior to spending from its FED account. Therefore excess reserves either exist OR MUST BE CREATED by the FED in order to support a treasury auction.
This is what Marrier Eccles is describing to congress in his 1947 testimony when he says,
"If Congress appropriates more money than Congress levies taxes to pay, then, there is naturally a deficit, and the Treasury is obligated to borrow. The fact that they cannot go directly to the Federal Reserve bank to borrow does not mean that they cannot go indirectly to the Federal Reserve bank, for the very reason that there is no limit to the amount that the Federal Reserve System can buy in the market. That is the way the war was financed.
"Therefore, if the Treasury has to finance a heavy deficit [borrow more reserves than there are in excess], the Reserve System creates the condition in the money market [by adding reserves] to enable the borrowing to be done, so that, in effect, the Reserve System indirectly finances the Treasury through the money market, and that is how the interest rates were stabilized as they were during the war, and as they will have to continue to be in the future.
"So it is an illusion to think that to eliminate or to restrict the direct borrowing privilege reduces the amount of deficit financing."
[] material/commentary added by me.
The purpose of Tsy issuance is to drain excess reserve to facilitate the Fed to hit its target rate. But what goes on in the FRS is much more complicated that this and to think of single transaction in aggregate such as a reserve drain is to conceptualize something is is actually included in a lot of other on-going operations, which is chiefly about the payment system and liquidity management. The Fed provides rb as needed for liquidity through lending, chiefly repos. The discount window is what corresponds in the FRS to an overdraft privilege on a customer deposit account at a bank. If a bank comes up short required reserves in a period, then the Fed automatically lends the funds at a penalty rate called the discount rate. Similarly, the PD's have accounts at the Fed and when they need liquid funds in the form of rb they just repo some tsys from their inventories. Sortin out how individual transactions occurs is not possible without access to the system, but that's OK, since what is important for macro analysis is what happens in aggregate. What happens in aggregate is that Tsy issuance offsets Tsy deficit expenditure.
@Ken,
"clearing banks" in this sense means an institution with a direct FED relationship. A "bank" may or may not have a direct relationship with the FED. Some small financial institutions still use correspondence clearing to handle their inter-bank settlement relationships (rely on clearing and settling payments with other banks/credit unions). Prior to the crisis (and the reason why Leheman failed - it didn't have access to the Discount Window) investment banks DID not have direct access to the FED or the Discount Window. You have to own a depository institution to have a FED depository account.
That said, whether or not you have a direct FED relationship, inter-bank clearing impacts demand for federal funds - the FED target rate.
Ken The following from the NY Fed ... since it talks about PD accounts at "clearing banks", they are obviously not assuming that the PD itself is a bank
Here's the FRBNY:
• The Fed uses repurchase agreements, also called "RPs" or "repos", to make collateralized loans to primary dealers. In a reverse repo or “RRP”, the Fed borrows money from primary dealers. The typical term of these operations is overnight, but the Fed can conduct these operations with terms out to 65 business days.
• The Fed uses these two types of transactions to offset temporary swings in bank reserves; a repo temporarily adds reserve balances to the banking system, while reverse repos temporarily drains balances from the system.
• Repos and reverse repos are conducted with primary dealers via auction. In a repo, dealers bid on borrowing money versus various types of general collateral. In a reverse repo, dealers offer interest rates at which they would lend money to the Fed versus the Fed’s Treasury general collateral, typically Treasury bills....
There are two main types of settlement methods for repos: triparty and “delivery vs payment” or DVP. Fed repos are done via triparty settlement, which means that the Fed and the primary dealers use a triparty agent to manage the collateral. In a triparty repo, both parties to the repo must have cash and collateral accounts at the same triparty agent, which is by definition also a clearing bank. The triparty agent will ensure that collateral pledged is sufficient and meets eligibility requirements, and all parties agree to use collateral prices supplied by the triparty agent.
Wikipedia-Triparty Agent:
The distinguishing feature of a tri-party repo is that a custodian bank or international clearing organization, the tri-party agent, acts as an intermediary between the two parties to the repo. The tri-party agent is responsible for the administration of the transaction including collateral allocation, marking to market, and substitution of collateral. In the US, the two principal tri-party agents are The Bank of New York Mellon andJP Morgan Chase.
Ok ... after going through everybody's comments, not sure what we disagree about really. I believe I understand how it all works, thanks to reading Warren Mosler's stuff.
The only reason I engaged in this discussion is to check my own understanding.
Understand repos/reverse-repos, and agree those are essentially loans between fed and PD and vice versa, ... the treasuries themselves being the collateral.
That is part of normal monetary operations and is quite a different thing than the temporary PDCF.
But the article's author wasn't talking about repos. He's talking about outright purchase of bonds by the PD from the Fed. He said:
"3. The Federal Reserve increases the reserves of the primary dealers to cover their money transfer to the government."
I still contend that this statement is not right. Why would the Fed want to "cover their money transfer" when this money transfer is the whole point of the operation (selling the bond)? Seems the Fed would **want** this financed through the money markets, not themselves, since they are **trying** to make a net withdraw of reserves from the system.
Yes, I suppose the PD could try to repo some treasuries it already holds to the Fed to obtain the funds to buy new ones from the Treasury. That would be one possilbe source of financing for the PD. But those would be independent operations ... it wouldn't be a case of the Fed supplying the funds so the PD can buy the bond. The Fed engages in these repo operations to hit its monetary targets ... not to specifically enable this other bond buying transaction by the PD.
"3. The Federal Reserve increases the reserves of the primary dealers to cover their money transfer to the government." I still contend that this statement is not right.
Agreed, I would guess this statement is operationally incorrect but someone else would have to confirm as I don't have direct PD operational knowledge.
"Why would the Fed want to "cover their money transfer" when this money transfer is the whole point of the operation (selling the bond)? Seems the Fed would **want** this financed through the money markets, not themselves, since they are **trying** to make a net withdraw of reserves from the system."
This still yields a system overdraft which would be INDIRECTLY covered by the FED. Going back to Marriners statement, the FED funds the expense either directly or indirectly - there is no way around it, smoke and mirrors aside.
Adam, if you're just pointing out that the Fed is always constrained to provide sufficient reserves to the system as a whole to support its target interest rate, I agree ... at least under a normal policy regime sustained by open market operations.
But in this particular instance, I still go back to the point that achieving a net decrease in reserves in the system is the whole point of a bond sale.
"3. The Federal Reserve increases the reserves of the primary dealers to cover their money transfer to the government."
I still contend that this statement is not right. Why would the Fed want to "cover their money transfer" when this money transfer is the whole point of the operation (selling the bond)? Seems the Fed would **want** this financed through the money markets, not themselves, since they are **trying** to make a net withdraw of reserves from the system.
The PD's are brokers who sell the tsys they purchase on. The Fed is justs providing liquidity to the PD's to facilitate the transaction with the private sector as the govt's fiscal agent acting in concert with the PD's a brokers. The transaction is actually between the Tsy and the ultimate purchasers of the Tsy security that provides the rb from excess rb. If the Fed lends to the PD to provide liquidity that is canceled out in the overall transaction.The PD's make a very small fee on each transaction as the market markers. Financial markets are intermediated to provide maximum liquidity at the lowest transaction cost.
But in this particular instance, I still go back to the point that achieving a net decrease in reserves in the system is the whole point of a bond sale.
This happens when the rb used by non-govt to purchase the tsys is credited to the TSy account. It's "to the penny" because the total amount of rb from the sale (asset on Tsy book) balances the total amount of tys issuance (liability on Tsy book). That's why this is "in aggregate."
"The Fed is justs providing liquidity to the PD's to facilitate the transaction with the private sector as the govt's fiscal agent acting in concert with the PD's a brokers."
Under what program/facility does the Fed provide this financing? I don't think it exists.
It's called repo. PD's use their inventory to borrow from Fed to purchase tsys which bear greater interest than the PD's pay the Fed since they borrow at the FFR, the lowest rate. Then they sell their stock on as the market makers.
Agree they do repos with each other. Again ... this is a mechanism for fine tuning monetary policy. Under normal interest rate regime and open market operations, repos and reverse repos are mostly used to maintain the policy rate. Whereas bond sales/purchases are used when a major move of the policy rate is desired. Or for unconventional operations like QE.
Repos are not a mechanism designed to finance the purchase of bonds from the gov't. If the PD happens to use it this way as part of their financing options, that would be incidental.
In any case, I don't think the author was trying to make some minor point about fine tuning monetary operations via repos. He wouldn't list this as one of his three main points if that were the case. He's trying to make some big picture points in his short article, and I think he has shown that he has a conceptual error.
It's called repo. PD's use their inventory to borrow from Fed to purchase tsys which bear greater interest than the PD's pay the Fed since they borrow at the FFR, the lowest rate.
What are the implications on that when the effective FFR is higher than the yield on the treasury (as it is now out to 1 year)?
What are the implications on that when the effective FFR is higher than the yield on the treasury (as it is now out to 1 year)?
Strong incentive to for Pd's not to hold inventory they have borrowed against and to sell on
Bob, the only way you are going to be satisfied is by asking John Lounsbury directly, it seems.
The recent schedule of Fed repo operations is available here:
http://www.newyorkfed.org/markets/omo/dmm/temp.cfm?SHOWMORE=TRUE
As you can see, they don't happen very often these days. And all of the ones that did occur were part of the "operational readiness program", not to actually affect reserve levels. This program is described here:
http://www.newyorkfed.org/markets/opolicy/operating_policy_101012.html
Note specifically this statement:
"The operations have been designed to have no material impact on the availability of reserves or on market rates."
Post a Comment