On Central Bank Lending to Government

The central bank prints money, lends it to the government, and the government sooner or later spends it (or uses it to cut taxes or increase transfer payments).

There seem to me to be two views on this question that are equally daft:

1. The Orthodox Daft View: "Central bank lending to government is a Bad Thing and should be prohibited!"
2. The Revolutionary Daft View: "Central bank lending to government is a Good Thing and should be mandatory!"

I don't get it.

First off, central bank lending to government is very orthodox and very traditional. The Bank of Canada is doing it now. The Bank of Canada owns government of Canada bonds. Does it really matter whether the Bank of Canada bought those bonds directly from the government, or indirectly from someone else (who could be anyone) who bought them from the government?

Second off, if the government of Canada owns the Bank of Canada (which it does), and if the Bank of Canada pays all its profits to the government (which it does), it does not matter what interest rate the Bank of Canada charges the government of Canada. If the Bank of Canada charges the government of Canada an extra gazillion dollars in interest, that simply means the Bank of Canada earns an extra gazillion dollars in profits, and the Bank of Canada pays an extra gazillion dollars to the government of Canada. It's a wash.

Third off, even if the Bank of Canada refused point blank to buy government of Canada bonds (either directly or indirectly), and bought commercial bonds instead, money is fungible. If the Bank of Canada switches from buying government bonds to buying commercial bonds, the people who would have bought commercial bonds now buy government bonds instead. Does it make a difference? Yes, it probably does, because the two sorts of bonds will not be perfect substitutes. But it's not like there's any Big Difference of Principle involved. The government now pays interest to the public, and the public now pays interest to the Bank of Canada, and the Bank of Canada pays that interest back to the government as profits.

If I lend money to Tom, and Tom lends money to Dick, and Dick lends money to Harry, am I indirectly lending money to Harry? We can't answer that question by putting a tracking device on the money I lend.

If the government of Canada owns the Bank of Canada, which it does, then de facto direct or indirect Bank of Canada lending to the government of Canada, at an irrelevant rate of interest, is neither a Bad Thing nor a Good Thing. It's inevitable, if the Bank of Canada lends and the government borrows.

What would make a difference is if we passed a law that said the Bank of Canada could only lend to the government of Canada, and the government of Canada could only borrow from the Bank of Canada. That law would tie monetary and fiscal policy together, like two runners in a three-legged race, so neither could change without changing the other. The fiscal deficit in any month would have to equal the amount of money printed in that same month. You could have an independent monetary policy, or you could have an independent fiscal policy, but not both. Either the amount of money printed would determine the fiscal deficit; or the fiscal deficit would determine the amount of money printed.

Such a law sounds daft to me. Why would you want to do that? How would adding an extra constraint improve monetary policy and/or fiscal policy? Mightn't there be times when you wanted to run a deficit, but wanted to reduce the stock of money? Or print money but wanted to run a surplus? In general, adding an extra constraint on choices leads to worse choices, because the constraint sometimes prevents you choosing something you would have chosen. Now there are exceptions to that general rule, like time-inconsistency in dynamic games, where adding a constraint on your own choices lets you change others' expectations of your future choices, which changes their current actions in ways that make you better off despite the extra constraint. But I don't know of any such argument for this particular constraint.

Any money the Bank of Canada prints lets it buy assets, which earn interest, which give the Bank of Canada profits, which it pays to the government, which the government eventually spends. Sooner or later, the government of Canada will spend any money printed by the Bank of Canada. It's inevitable. But the "sooner or later" bit is very important. It doesn't have to be this month, or even this year, or even this decade. Nor should it be.

And the ECB is a bit weird, because it isn't really owned by any single government, so maybe things are different in the Eurozone.

Comments

[JKH]

The Eurozone is very different - because the original design and intention was that the ECB should lend to banks instead of government(s).

[Nick Rowe]

JKH: OK. But why? Should we pass a similar law in Canada?

[Nick Rowe]

And if the ECB lends to banks, and the banks lend to governments, does that intention even make sense?

[JKH]

because the Eurosystem was ultra-anti-inflationary in original design

not allowing backdoor "monetization" of government debt

Eurozone governments get the benefit of seigniorage (though similar net income allocation as done by other central banks) - but not the net benefit of funding

I think its an unnecessary rigidity in the case of Canada

[JKH]

Eurozone bank lending to governments was voluntary - not structural

[Nick Rowe]

JKH: Again, why does it make a difference whether the central bank buys government IOUs, or the central bank buys bank IOUs and the banks buy government IOUs? And why does it make a difference in the ECB case but not in the Canadian case?

[Nick Rowe]

Now there is one difference I can see, between Canada and the ECB. Because we have one government that gets BoC profits and they have several that get ECB profits. If the BoC buys GoC bonds at an above market price, and earns lower profits, it's a wash for the GoC. But if the ECB buys Greek govt bonds at an above market price, and earns lower profits, it is not a wash for the government of Greece, which only gets a small share of those profits.

[JKH]

Because in the Eurozone, its a private sector choice ("normally") as to what the sources of government funding are. Could be any of banks, pension funds, households, etc.

There's no structural "monetization" of government debt by the central bank.

[Brent Buckner]

Nick Rowe:
I think it makes a difference in the ECB case because in that case a government that gives an IOU is not guaranteed to receive the (interest) proceeds - Greece may issue an IOU for which it is fully responsible, but Greece does not fully own the ECB.

[louis]

So is this a response to Bernanke's Helicopter Money post? It seems he (having seen the political, cultural and technical constraints within which the Fed operates) has come up with a convoluted mechanism for demonstrating to markets that a monetary expansion is permanent / letting the monetary authority instigate countercyclical fiscal policy.
In your framework, price level path targeting (to which he nods) may do the heavy lifting, and it may be completely unnecessary for the Fed to lend money directly to the Treasury, vs buying some asset in the open market.
Still, is there any situation where this sort of arrangement makes sense to you?

[JKH]

Would have been interesting to see how the Eurozone crisis would have evolved if the ECB had been mandated to acquire pro rata shares of Eurozone debt from the outset instead of lending to banks.

[Nick Rowe]

Brent: "I think it makes a difference in the ECB case because in that case a government that gives an IOU is not guaranteed to receive the (interest) proceeds - Greece may issue an IOU for which it is fully responsible, but Greece does not fully own the ECB."

That's the direction I'm leaning too. (See my 9.20 comment, we were probably writing at the same time.) But it's still not totally obvious, and depends on how much the ECB can move the market price for Greek bonds relative to other government bonds, and whether the Greeks can do a Trojan Horse job at getting control of the ECB.

[Nick Rowe]

louis: not directly related to Bernanke's post. The immediate trigger was Tom Warner's most recent comment on my Helicopter Bonds post. Plus a bunch of Canadian lefties recently put out a policy statement in favour of the Revolutionary Daft view.

[Nick Rowe]

JKH: the buying the pro rata share (or buying the index) policy does seem to be a way to get around the risk of favouritism, when the central bank is buying something other than the bonds of the government that owns it 100%.

[JKH]

a pro rata purchase policy could have been done on the same basis as the profit allocation policy - a share according to what they call "capital keys"

[Tom Warner]

For most of the world the ban on direct central bank lending to government is an antiquated taboo stemming from the history of the gold standard. Central bank lending to government was one way to bypass the constraints of the gold standard on fiscal policy, and it tended to end very heinously, eg Germany 1920s. Tightening the discipline of the gold standard by rigorously separating monetary policy from fiscal policy and banning central bank lending to government was all about avoiding inflation.

In the gold standard, when the government borrowed from the private sector, it committed to pay gold. And when the governemt borrowed from the central bank, it was just creating gold certificates without adding to its stock of gold. Banning that lending was all about stopping government from handing out multiple claims on its gold.

In the fiat currency era, there is no gold standard stricture. Thus in most countries central bank lending to government does not really bypass any stricture, and there's not really any logic to banning it. Instead independent inflation-targeting monetary policy plays the role of avoiding inflation. And those countries that lack it or do it half-heartedly do indeed have much higher inflation.

The exception is the Euro Area, where euros represent a liability of the entire Euro Area, but are issued by national central banks. Thus any one country's national central bank could by lending to its government increase its country's claims on the wealth of the whole Euro Area. The ban on monetary financing in the Euro Area is all about that problem, which is unique to an international currency zone.

[Tom Warner]

It might seem not that important in a gold standard whether a government issues bonds linked to gold or hands out certificates denominated in gold. But excessive bond issuance affected a narrow group of lenders, usually by default, and thus they could contain it by refusing to lend to government, whereas excessive currency issuance affected the broad population through inflation.

[Nick Rowe]

Tom: I can see how making the central bank hold 100% gold reserves against currency, so it is banned from lending to anyone, would make a difference. But I can't see the point of banning the central bank from lending to the government, but letting it lend to the private sector.

[Too Much Fed]

"Does it really matter whether the Bank of Canada bought those bonds directly from the government, or indirectly from someone else (who could be anyone) who bought them from the government?"

I thought the idea was that the market should set the interest rate of new bonds at auction.

"Second off, if the government of Canada owns the Bank of Canada (which it does), and if the Bank of Canada pays all its profits to the government (which it does), it does not matter what interest rate the Bank of Canada charges the government of Canada. If the Bank of Canada charges the government of Canada an extra gazillion dollars in interest, that simply means the Bank of Canada earns an extra gazillion dollars in profits, and the Bank of Canada pays an extra gazillion dollars to the government of Canada. It's a wash."

What if there are so many central bank reserves that pay interest that all of the interest from gov't bonds goes to the commercial banks (the central bank breaks even with no profits)?

[Jacques René Giguère]

Tom : whether the Reichbank had or not lent to the Weimar gov't in fall 1923 was immaterial. The Franco-Belgian occupation of the Ruhr and the reaction to it made an economic catastrophe inevitable. But that story later provided a good cover for the elite ultimate responsibility in bringing He-Who-Must-Not-Be-Named to power.

[Majromax]

@Too Much Fed:

> What if there are so many central bank reserves that pay interest that all of the interest from gov't bonds goes to the commercial banks (the central bank breaks even with no profits)?

The interest rate on government debt held by the central bank still doesn't matter; reserve balances will increase by X% regardless, which will ultimately be removed from the seigniorage revenues remitted to the central government. Adding $I interest to the government->central bank->government loop does not change that in any way.

@Nick Rowe:

> But I can't see the point of banning the central bank from lending to the government, but letting it lend to the private sector.

I think the Eurozone has the effect nailed down: if the central bank does not hold private-sector debt corresponding to the entirety of the public debt, this on-paper restriction forces the government to act like it can borrow only from the private sector. That makes its debt subject to a risk premium.

I'm not sure why this would possibly be considered a good thing, but it does have an effect.

[Tel]

Debt is a promise, and promises are either fulfilled or broken. Government bonds are merely a different type of debt instrument. Most of our money is backed by debt (i.e. promises) one way or another.

That law would tie monetary and fiscal policy together, like two runners in a three-legged race, so neither could change without changing the other. The fiscal deficit in any month would have to equal the amount of money printed in that same month.

That's pretty much inevitable, once you understand that banks use government bonds as "high quality" reserves backing new debt creation. So there might be a handful of the "mom and pop" investors who forgo consumption, buy a government bond, and then hold that bond to maturity... in that case government deficit is not creating new money. However when the bond is owned by a private bank, that's a different matter and you are back in the three legged race.

To put this in a slightly different way: government bonds are adequately close to a money substitute, therefore printing bonds is always quite similar to printing money.

Mind you, banks may choose not to lend, and just hold the reserve like the mom and pop. Or government can attempt to use regulations to force banks to hold more reserves (like China did a few times). These actions are not directly linked to deficit spending.

Mightn't there be times when you wanted to run a deficit, but wanted to reduce the stock of money?

In other words, transfer private debt over to public debt. I guess my feelings on such things probably go without saying. Tends to happen when bank debt is unwinding (i.e. debt deflation, defaults on bad debt, etc) and government attempts to compensate with big spending projects to prop up certain politically protected entities during the time of difficulty.

[JF]

Bank of Canada 'lends' to the govt???? When govt revenues from its ways and means are less than needed ti cover claims the govt sells a debt instrument to the public, extracting cash from the private buyers of the bonds. Combined this financing shows cash being managed so claims are paid (normally with a bit of cash still in the accounts. The govt does NOT borrow the money from the central bank of Canada at any time in this telling of how it works - unless in Canada do they borrow the money from their own agency (because this agency is permitted to create bond and create money too, I guess.

I dont think this story has been told correctly. The central bank does not give bonds to the govt do they, and how does this get a claim paid?

[Henry]

"That law would tie monetary and fiscal policy together"

This may limit how much money the central bank can print, but the bank still set interest rates.

[Frank Restly]

Nick,

"Does it really matter whether the Bank of Canada bought those bonds directly from the government, or indirectly from someone else (who could be anyone) who bought them from the government?"

From a political economy sense, yes it does matter.

Government bond auctions should be permitted to fail as a means of expressing disproval over government policy choices. If the federal government does something you find abhorrent and utilizes borrowed money to fund such policies would it make a difference to you if they were borrowing money at 1% interest or 20% interest? This is why having the central bank lend money directly to the government is a bad idea. In a sense, it short circuits the political process.

[Too Much Fed]

JKH, does new gov't bond vs. existing gov't bond matter here?

[Too Much Fed]

Majromax said: "The interest rate on government debt held by the central bank still doesn't matter; reserve balances will increase by X% regardless, which will ultimately be removed from the seigniorage revenues remitted to the central government. Adding $I interest to the government->central bank->government loop does not change that in any way."

Can you give me some numbers? I am not getting that at all.

One other thing. I am assuming that all interest payments are made using existing "money".

[John Handley]

Regarding the ECB (and moving the topic very slightly to FTPL also pardon me if this has already been somewhat discussed, I didn't bother to read the comments above), wouldn't the only difference between a normal country and the Euro Area be the number of different government bonds in the GBC? If you have one currency area and ten different countries, then the GBC would be M_t + B_t(0) + B_t(1) + ... + B_t(10) + T_t(0) + T_t(1) + ... + T_t(10) = R_t-1(0) B_t-1(0) + ... + R_t-1(10)B_t-1(10) + M_t-1

In this case, what happens when only one of the governments, e.g. Greece, is non-Ricardian while the other governments are Ricardian? I guess empirically the ECB wins the game of chicken and Greece becomes Ricardian, but what happens if the troika doesn't have the political power to force Greece to become Ricardian? Does the ECB end up losing control of inflation?

[Tom Warner]

Nick: The point is to get your head back in the late 19th and early 20th century when these norms were established, and think as people thought then. It made sense then to ban central bank lending to government, because that was associated with currency failures.

I'm surely not saying they had it all figured out back in any good old days. They had inflation and deflation, boom and bust cycles despite the gold standard, until they decided it wasn't helping bring stability after all.

The main reason why it can make sense for central banks to lend to "the private sector" but not be allowed to lend to government is so they can fill their primary function as lenders (to banks) of last resort. They usually aren't allowed to or just by tradition do not lend to the real private sector, because that's considered a fiscal role that should be reserved to elected authorities.

Sure, CB lending to banks allows them to dilute currency in exactly the same way that lending to government can. That has never been a cause of hyperinflation, since central banks don't become captives to private banks to that extreme degree. But it has been a common cause of chronic high inflation and remains so among EMs. And the answer to it which many of those EMs still lack is rigorous inflation targeting.

[Nick Rowe]

Majro: not sure I am following you there. If the national debt is bigger than the central bank, then the marginal lender to the government is always a private lender?

Tel: Not sure I'm following you either. But one empirical observation: in the mid 1990's the Canadian government tightened fiscal policy a lot (for reasons unrelated to reducing aggregate demand), but the Bank of Canada managed to offset that fiscal tightening and keep inflation on target.

JF: I thought I had already answered that in the post. If the government sells a bond to Tom, and Tom in turn sells that bond to the Bank of Canada, all that means is that Tom is acting as middleman. Would it make any difference if the Bank of Canada bought the same bond directly from the government?

Henry: you might be onto something there. If the CB can vary both the quantity of money *and* the rate of interest paid to people who hold that money, that gives the CB an extra degree of freedom. But the Bank of Canada's profits come mainly from its monopoly on currency, which pays a fixed 0% interest (nominal).

Frank: you will always find a middleman willing to buy government bonds if he know he can resell them to the central bank.

John: Yep. That's the danger of a central bank acting as *unconditional* lender of last resort, whether to a commercial bank or to a government. Suppose the Bank of Montreal (BMO) gets into trouble, and the Bank of Canada steps in as LOLR. If it does so unconditionally (as opposed to ordering BMO to get its act together), then the BoC is now responsible for maintaining the fixed exchange rate between BMO dollars and BoC dollars. Alpha and beta have changed places. BMO is now the new central bank, and can create as many BMO dollars as it likes, knowing the BoC must follow along.

Tom: maybe. It's presumably something like that, that seemed to make sense at the time.

[Scott Sumner]

Nick, Good post, and I think some of your commenters are missing the big picture. I've done thought experiments where the central bank stops injecting base money through bond purchases, and begins paying public sector salaries with new base money (cash). This causes the Treasury to issue fewer bonds, so that the net holdings of bonds by the (non-Fed) public is identical to what it would be with ordinary OMOs. It's all a wash, at least as a first approximation.

[JF]

Nick you referred to the central bank, an agency of the govt lending money to the other agencies. It was the use of the word 'lend' that seemed to be a misspeak, but I thought I should ask. Maybe that is how it works in Canada.

In the US, by law, the Fed cannot buy a Treasury debt-security directly from the govt, it must buy it first from a private buyer who would have bought the bill or bond in a prior period. This I consider to be anachronistic though it is a way to make the public's govt dependent on private financial system interests - a political result, blatent rent seeking result of the powerful who dreamed this up initially (while they held unlimited power to create credit/money, their endogenous privilege, but they stopped the ability of the public to do it by establishing the rule or law of finance).

As to your question, my answer is that the central bank should have the discretion to buy bonds direct from the govt. It avoids the control and subsidization of turnstiling the bonds thru the private sector, just is administratively more efficient, while the discretion itself adds a powerful monetary discipline tool to keep the primary dealers and the financial asset trading marketplaces on alert since they depend NOW on using govt bonds as a medium of exchange, and account (to go with their protection of the principal). Big difference in terms of subsidy and power but no difference from a cash perspective so the law needs to change to permit the discretion to buy direct from the govt. I've asked Bernanke to write about this (and about redemption of the bond as a monetary tool).

[JF]

Forgot an important point. When a bond is sold by the govt private interests use earned cash to make the purchase, this extracts money off of their books away from other opportunites (while the govt turnstiles the cash back into the economy as it pays claims). There is a contractionary aspect to this at least in a top-down investor view of the world (perhaps the term 'crowding out' captures the shift away from one private interest giving up on their opportunity to use this cash in other ways). The ability of the central bank to buy a bond directly from the govt would avoid any contractionary effects while the new money used to pay claims clearly increases the money supply which may help during downturns ( when this helicoptering mechanism should be considered for use to some degree).

[Majromax]

@Nick Rowe:

> If the national debt is bigger than the central bank, then the marginal lender to the government is always a private lender?

Right. Imagine a Eurozone where the ECB was required to buy government bonds on a per-capita or per-GDP basis, and it returned its seiginorage revenue on the same basis.

Now, some enterprising country decides to eliminate its national debt, such that it issued only the debt that the ECB was required to buy. That nation would be a price-setter, and it could use that power to extract seigniorage revenue from the ECB at the expense of other nations.

@Too Much Fed:

> Can you give me some numbers? I am not getting that at all.

Let's attach numbers: bank reserves are $1bn, the interest rate on reserves (and bonds) is 10%, and we'll vary the stock of bonds held by the central bank.

If the central bank holds $2bn of debt, in each period the government pays $200m to the central bank, the central bank pays $100m to the banks, and the net seigniorage revenue is $100m that is remitted to the government. The central government faces a let loss of $100m on (debt servicing less seigniorage).

If the central bank holds $1bn of debt, in each period the government pays $100m interest to the central bank, the central bank pays $100m in IOR, and the net seignorage revenue is $0. On the balance, the fiscal authority still faces a net loss of $100m.

If the central bank holds no government debt, in each period the government pays $0m interest to the central bank, the central bank pays $100m in IOR, and the central bank is in the hole by $100m. Since a bankrupt central bank sounds like a stupid idea, the government has to bail it out to the tune of $100m, which is the net loss to the fiscal authority.

@JF:

> the Fed cannot buy a Treasury debt-security directly from the govt, it must buy it first from a private buyer who would have bought the bill or bond in a prior period. This I consider to be anachronistic though it is a way to make the public's govt dependent on private financial system interests

This is an empirical question. Does the US Fed, when engaging in open-market purchases, consistently purchase bonds at a higher price than the auction price? If so, then this is a quantitative measure of the rent extracted; if not then the activity is little more than an accounting exercise.

[Frank Restly]

Nick,

"You will always find a middleman willing to buy government bonds if he knows he can resell them to the central bank."

The would depend on the profitability of the trade. How many middle men will buy a government bond for $10 and sell it to the central bank for $5?

Also, I said "government bond auctions should be allowed to fail" as a policy recommendation. I realize that most countries have enacted policies to prevent such an event - for instance central bank lender of last resort function, central bank open market operations, and primary dealer arrangements.

This gets back to your statement:

"1. The Orthodox Daft View: Central bank lending to government is a Bad Thing and should be prohibited!"

I don't consider this a daft view. I consider the decision to purchase / not purchase government bonds both a political and an economic decision.

[Jussi]

"some enterprising country decides to eliminate its national debt"

This gets it backwards. A big price insensitive buyer (the ECB) doesn't mean that there is an incentive to sell less.

[Nick Rowe]

Scott: thanks. Your thought-experiment makes sense to me. Aside from central bank independence (*who* gets to decide how much money gets printed), which is obviously important, we might as well consolidate the government and central bank balance sheets.

JF: "In the US, by law, the Fed cannot buy a Treasury debt-security directly from the govt, it must buy it first from a private buyer who would have bought the bill or bond in a prior period."

Right. Given a liquid enough and competitive enough market in bonds, that silly law doesn't make any difference, except maybe an extra bit of hassle, and maybe extra commissions.

If there is a short delay between the government borrowing money and spending that money, some amount of money will be withdrawn from circulation during that short delay. But I don't think that matters much, since the central bank can take offsetting action.

Majro: "Now, some enterprising country decides to eliminate its national debt, such that it issued only the debt that the ECB was required to buy. That nation would be a price-setter, and it could use that power to extract seigniorage revenue from the ECB at the expense of other nations."

Good point. Not sure how realistic it is though.

[Nick Rowe]

Jussi: I think Majro is right. Suppose the ECB is required to buy $1 billion worth of my debt each year. So each year I issue one nearly worthless bond that pays 1 cent per year forever, and the ECB must pay me $1 billion for it.

[Too Much Fed]

Majromax said: "If the central bank holds $2bn of debt, in each period the government pays $200m to the central bank, the central bank pays $100m to the banks, and the net seigniorage revenue is $100m that is remitted to the government. The central government faces a let loss of $100m on (debt servicing less seigniorage).

If the central bank holds $1bn of debt, in each period the government pays $100m interest to the central bank, the central bank pays $100m in IOR, and the net seignorage revenue is $0. On the balance, the fiscal authority still faces a net loss of $100m.

If the central bank holds no government debt, in each period the government pays $0m interest to the central bank, the central bank pays $100m in IOR, and the central bank is in the hole by $100m. Since a bankrupt central bank sounds like a stupid idea, the government has to bail it out to the tune of $100m, which is the net loss to the fiscal authority."

Those sound good. Also, those do not appear to be a "wash" to me.

Next and pre-2008 in the USA, the fed did not buy new gov't bonds and did not do QE. So, it seems to me the fed was reacting to desired demand for currency, desired demand for central bank reserves, and required demand for central bank reserves while keeping the fed funds rate on target. The fed was "lending" or doing an OMO for currency and/or central bank reserves to the commercial banks with the gov't bonds as collateral/outright possession to back the currency/central bank reserves, not lending to the gov't.

To see that, assume the reserve requirement was 0, no entity wanted currency, and the commercial banks did not want any central bank reserves. Entities turn in currency for demand deposits, the commercial banks turn in all of the currency for central bank reserves, and the commercial banks turn in all of the central bank reserves for bonds.

[Too Much Fed]

Nick's post to John: "John: Yep. That's the danger of a central bank acting as *unconditional* lender of last resort, whether to a commercial bank or to a government. Suppose the Bank of Montreal (BMO) gets into trouble, and the Bank of Canada steps in as LOLR. If it does so unconditionally (as opposed to ordering BMO to get its act together), then the BoC is now responsible for maintaining the fixed exchange rate between BMO dollars and BoC dollars. Alpha and beta have changed places. BMO is now the new central bank, and can create as many BMO dollars as it likes, knowing the BoC must follow along."

John, what if getting BMO's act together means solvency and nothing else?

[JF]

@majromax - "Does the US Fed, when engaging in open-market purchases, consistently purchase bonds at a higher price than the auction price?"

I look more to the coupon being paid by the US Treasury compared to Germany; why is it that Germany's bonds provide so little when that is a depopulating energy-dependent country tied to an economic mess with financial and political uncertainty at this time? There is a subsidy when looked at this way. When a primary dealer keeps some bonds and sells them in the open market rather than to the FED they will charge more to the second buyer than the FED pays them, and this pricing control stems from the position accorded them, another subsidy to the dealer/bank. Interest on reserves, which is positive, is another subsidy. To me the subsidy is not apparent in just a particular auction. If the FED were able to buy direct instead they would buy a mix of instruments at a settled price intended to lower the cost to govt and remove some of this subsidy, and the threat that this would be so in unknown amounts would bring discipline to this marketplace. Perhaps the Fed could simply say we'll 'buy' as ,uch as you want to issue and we are happy with an expected return just as Germany is getting.

@Nick - There really isn't a delay as the cash position of the Treasury accounts is static, and it is a miniscule amount. I look at the QE purchases in the hundreds of billions. I can't explain where the money/positions were prior to 2008 (there were little to no excess reserves in this 'officialdom' system) but we now see reserves lay there statically and I assume there is a contractionary effect indicated by this, the sums are huge/enormous. Definitely missing something, but I am most concerned that the later QE buying programs are in fact contractionary of the lending capacity of the banking system and the reserve positions are indicative of money not reaching the real economy. Plainly, buying direct with new money that is immediately outlaid by the govt, especially with an agreement to redeem the principal via an offset across the two sets of govt-agencies' books (essentially you erase the positions) is classical pump priming (via the outlaying) avoiding any of this banking system mess, costs, delays (maturities may delay some effects), unnecessary subsidies.

But what really fries me is that the silly law as you refer to it, really walls off the sovereignty and makes it (yes, the people) dependent on the financiers to get cash to pay lawful claims. That is wrong. And it is an anachronism, it ain't Hamilton's colonial debt era.

[Antti Jokinen]

Tom Warner wrote (under another post, but this subject is not that far from it):

"Currency is a liability similar to an eternal bond that never has to be redeemed."

If this is what currency is, then what sense does it make to say that a (non-eternal) bond is redeemed by handing over currency to the bondholder (creditor)? All that happens here is that a non-eternal bond is replaced by an eternal bond. If you call that "a redemption", then you might as well call stealing "buying".

Is there a paradox here?

[Nick Rowe]

Antti: well, if there were an *unanticipated* inflation, you might indeed say that the bond wasn't really (fully) "redeemed". There's an implied or explicit (2% inflation target) promise with money just as there is an explicit promise with bonds. The latter depends on the former (unless it's an indexed bond).

The paradox is that paper money is worthless if you have to keep it in your pocket forever.

[Antti Jokinen]

Nick, I don't see what inflation has to do with this. Unanticipated inflation or not, how do we defend our choice of words here? Correct me if I'm wrong, but in general we talk about redemption of a debt/liability (a bond is a special case). And I'm with Tom here: Currency indeed is, first of all, a liability, and second, it is similar to a bond that never has to be redeemed (but might very well be redeemed, and is probably even expected to be redeemed at one point). A bondholder holds a claim, and if you hand currency over to him, he still holds a claim. What was redeemed? A promise to replace a claim with another kind of claim? If so, I might say that a promise was *fulfilled*, but there is no *redemption* in any meaningful sense of the word. (Of course I understand how we are used to think, and say, that money can be used to pay back debt; it is this very convention of ours I'm questioning here.)

I don't want to sound like a crank. Behind this is a constructive aim. But you might think I'm nuts if I spell it out here, so I rather not do it, yet.

[Nick Rowe]

Antti: paper money is a very peculiar sort of "liability". A liability is some sort of promise. In the case of currency, the BoC that issued it promises: "I will try to make this bit of paper depreciate at around 2% per year against the CPI basket of goods".

[Antti Jokinen]

Nick, why this sloppy language? "Liability", not a liability, and "some sort of promise"?

Let me present an unconventional -- yet operational and internally consistent -- perspective on how a credit in the CB book ("paper money"/reserves; this is accounting, all the same) behaves.

I would like to avoid calling this credit (something you find on the right-hand side of the CB balance sheet) a liability. It is part of a claim/liability pair (these are fungible, so we cannot point to a certain pair -- I'll get back to this below), but it is actually the claim part of it. Anyone who has a credit in the CB book, has a claim against something or someone. Now, according to the conventional perspective, this claim -- the credit -- is at the same time someone else's liability. And this someone else is the CB (again, according to the conventional perspective). If you say that it is a "liability", not a liability, I agree with you.

But I don't stop there. I find the liabilities (debits) which correspond (debit=credit) with all these claims (credits) on the left-hand side of the CB balance sheet. They look like genuine liabilities *of someone* to me. As I said, we cannot form pairs out of certain claims and liabilities, but in aggregate there is always a liability for every claim.

Try to view the CB as a bookkeeper, who itself doesn't have liabilities to anyone, nor claims against anyone. It might be hard to see the CB in this way, but like I said, this perspective is operational and internally consistent (I know because I've been working on it for a long time now, testing it and trying to falsify it). When the CB buys Treasury debt from a commercial bank, the claim the bank held against Treasury is switched to a more general claim against anyone or anything you can find on the left-hand side of the CB balance sheet. There is an increase in these general claims (we can call this "base money supply") and this increase corresponds with an increase in the left-hand side of the CB balance sheet where we find the something or someone these claims are against (corresponding liabilities). In this particular case, there is an increase in Treasury liabilities/debt on the CB books.

How does this sound?

[Nick Rowe]

Antti: I'm fine with that perspective. Simplest way to put it: irredeemable paper money is an asset of the person who holds it, but not a liability of anyone. It's net wealth. The only reason it appears as a "liability" on the central bank's balance sheet is that the accountants would get anxious otherwise. It's a familiar idea to me. The Pigou effect, and helicopter money, are based on it. Willem Buiter's the guy who explored it in most depth. I've done a few related posts, over the years. Here's one.

[Antti Jokinen]

No, Nick. You're not fine with this perspective (probably I didn't explain it carefully enough). This perspective says that "base money" is not net wealth. And this is because every time the CB issues "base money", there is a corresponding addition to liabilities in CB books (not to CB's liabilities, but to someone else's liabilities which we find on the left-hand side of CB balance sheet; alternatively there can be a decrease in another (non-base money) account which shows on the right-hand side of the balance sheet). As I said, the CB can be viewed as a central bookkeeper/accountant who keeps track of *other entities'* claims and liabilities -- not its own.

I appreciate Willem Buiter's expertise. He's very good. But even he doesn't get this. (Have you read his piece on "numerairology"? He's onto something important, but then misses it because he cannot see that USD or CAD are nothing but abstract units of account -- numerairés -- too. We live in a "pure credit economy", something, for instance, John Hicks expected will be the reality in the future, missing the fact that we were there already when he said so.)

P.S. I'm an accountant.

[Eric Lonergan]

Another interesting post. The EZ is different. In the EZ monetary policy dominates fiscal policy. The ECB can mandate that governments supply it with bonds, the consequences of its policies can add or subtract from budget balances, and it can force a government into default! The textbook assumption of fiscal primacy is an institutional contingency.

[Oliver]

Antti, you might enjoy this:

https://en.wikipedia.org/wiki/Quantum_economics#Money

[Tom Warner]

Antti: To quote myself more fully, I said that on a central bank's balance sheet currency is like an eternal bond that never has to be redeemed, but on the balance sheet of consolidated public authorities (government + central bank), currency is a liability redeemed whenever tendered for any private debt to the public sector. "Legal tender for all debts public and private" after all has important meaning.

[Tom Warner]

PS to Antti: I get what you're saying about the two sides of the CB balance sheet. So long as the asset side is sovereign debt, the point is the same as the one I just made.

Another way of talking about it is that fiat currency backed by sovereign debt represents financialization of the government's future revenues, and ultimately of the republic's future output. So do sovereign bonds. However the non-base part of broad money is financialization of whatever commercial banks lend against - property and equipment and personal and organizational future incomes.

When you look at money that way the ground falls out from under money quantity theories of inflation.

[Jussi]

I agree with Tom and Antti. Think about a very small economy, eg. 3-persons. Then it is IMO relatively easy to see how one of them owning credit money means necessarily a liability for the other two. One can settle the liability with two bilateral transactions. It is definately not net wealth.

[Antti Jokinen]

Oliver: Thanks! I'm aware of the clear similarities between Quantum Economics (and also MMT/Circuitism) and the viewpoint I've adopted. This might sound weird, but by adopting a pure "double entry bookkeeping view" of the monetary system, I've managed to get rid of money (other than an abstract unit of account type) altoghether -- and with that goes the possibility to make a *payment* by transferring money to a creditor/seller. That's where I differ a great deal from others. The underlying logic isn't necessarily that different from "chartalist" expositions of money/credit (following Mitchell-Innes), but radical changes in language go with radical changes in how one sees the system, and I seem to have arrived at a relatively simple theory. When you manage to explain money away instead of needing to answer "What money is?", simplicity follows. (I'll go through the details in a thought experiment I'm going to publish later in 2016.)

Tom: I hear what you say. I just put these things in a different way. I think it's important to realize that one who holds a claim (a credit balance in any bank's books) is always a creditor. This is of course self-evident, but why do we then say that one gets *paid* when one receives this kind of claim? I cannot reconcile the state of having got paid with the state of being, or even becoming (through a so called payment) a creditor. (This is not just semantics, as I've build a whole theory starting from the premise that one cannot make payments in "money".)

Related to this, I think it's confusing to talk about a debtor accepting his own liabilities (claims on him) as a payment, as for instance Randall Wray puts it. If you and I have verbally agreed that I owe you a beer, it is crystal clear to me that it is me who is going to make a (debt re-)payment by buying you a beer, and it would make very little sense to say that you pay for the beer by letting me off the hook. Likewise, the government is not a debtor to me -- the taxpayer is. If the taxpayer has a tax obligation of $100 this month, then he can sell something to acquire a credit worth $100. At this point we can say that he has already fulfilled his obligation, and we can view what happens next (the so called "tax payment") as just the part where the books get finally updated. The "monetary" system is only about keeping track of real world debt relations -- claims and liabilities; credits and debits -- expressed in an abstract unit of account.

Of course it's slightly more complicated than this, because we need to get our heads around the fact that if I hold a general claim (say, a credit in the CB books), I cannot point to any separate entity against which I hold a claim. And I've found out that it is close to impossible for some people to realize that if they hold a claim with a nominal value of $50, it doesn't mean that they are to receive "fifty follars" -- it means that they are to receive goods or services worth (priced at) $50.

This short exposition no doubt lacks in clarity, so I'm happy to elaborate should you have any questions.

You might want to check my comment to Nick's earlier post on helicopter money (I criticize Quantity Theory in it): http://worthwhile.typepad.com/worthwhile_canadian_initi/2016/04/helicopter-money-is-permanent.html?cid=6a00d83451688169e201bb08d6d1d3970d#comment-6a00d83451688169e201bb08d6d1d3970d

[Too Much Fed]

Tom and Antti, how do you think the lender of last resort works?

[Tom Warner]

Antti: We seem to see things mostly the same way, only using different terminologies. When I say money is financialization of future government revenues, and thus ultimately of the future income of the republic, I think I mean the same thing you mean when you say the government isn't the creditor, the taxpayer is. But I'm fine using the standard terminology that government is a creditor even if it is really only intermediating, just as I'm fine with the standard lingo that a bank is a creditor when it lends, even if it is only intermediating.

I can't agree though that one can't make payments in money. That makes no sense to me at all. If I gave you anything whatsoever that you reckon as valuable, in exchange for something that you gave me in return, or to settle a debt, then I paid you, period. Whatever it was I gave you is irrelevant.

[Antti Jokinen]

Tom: Yes, we are definitely not that far apart. But fortunately there is something we disagree on (otherwise it would be a waste of time to continue the discussion). You said: "Whatever it was I gave you is irrelevant." Let's take a chartalist approach, in an economy without a central bank. You sell a horse to the government. The government representative who you deal with gives you a Treasury note (or tally stick, or similar) in return. Don't you agree if I say that the note is a *promise to pay* (which the government issued on behalf of the people)?

Whatever you give me, if I deem it to be of sufficient value, I recognize as a payment -- but not a promise to pay later. For me, it would be illogical to think otherwise.

I'm not fine with the standard lingo. If a bank doesn't strictly speaking *lend* money, then we must focus on what it strictly speaking does, and forget these metaphors (which we learn to use already as children, and that makes it so hard to get rid of them). If we manage to do so, clarity follows -- eventually.

[Alex Godofsky]

Nick:

Henry: you might be onto something there. If the CB can vary both the quantity of money *and* the rate of interest paid to people who hold that money, that gives the CB an extra degree of freedom. But the Bank of Canada's profits come mainly from its monopoly on currency, which pays a fixed 0% interest (nominal).

Does it really derive the majority of its profits from the monopoly on currency (which I take to mean physical currency)?

It seems to me that electronic reserves are much more significant than currency.

[Ricardo Martins]

Nick (sorry for the late question)
The two examples you just showed are different in my opinion.
If the CB lends directly to the Government it's like Helicopter Money, but if it lends to the commercial banks for them to lend to the Government is like QE.
QE is like an OMO like you previously said, so per se, does not increase the money supply.
HM is newly printed money going to the real economy, so it will raise the money supply.
About the profits I think you are right both ways they will flow to the government, but the effects on the economy are different.
Tell me please if I am wrong

[Nick Rowe]

Alex: I looked at recent Bank of Canada balance sheets. On the liability side, there's about:

$70 billion currency (paying 0% interest)

$20 billion in Government chequing account (which is a wash, because government owns the Bank of Canada)

$0.5 billion members of Canadian Payments Association (banks?) reserves (paying 0.25% below overnight rate of interest)

So yes, physical currency is the big one.

Ricardo: "QE is like an OMO like you previously said, so per se, does not increase the money supply."

Both QE and OMO (same thing) *do* increase the money supply.

[Ricardo Martins]

An OMO increases the monetary base and increases the money supply by reducing interest rates.
My point was, if OMO and/or QE doesn't reduce interest rates (ZLB or ELB or whatever..) does it increase the money supply?
Thank you very much Professor
(Please read my blog, I would much appreciate!)

[Too Much Fed]

Nick said: "John: Yep. That's the danger of a central bank acting as *unconditional* lender of last resort, whether to a commercial bank or to a government. Suppose the Bank of Montreal (BMO) gets into trouble, and the Bank of Canada steps in as LOLR. If it does so unconditionally (as opposed to ordering BMO to get its act together), then the BoC is now responsible for maintaining the fixed exchange rate between BMO dollars and BoC dollars. Alpha and beta have changed places. BMO is now the new central bank, and can create as many BMO dollars as it likes, knowing the BoC must follow along."

So, you are saying the price inflation target applies to the lender of last resort function, right?