Alpha, Beta, ECB independence, and Omega lender of last resort

There is a tension between central bank independence and acting as lender of last resort. We need to recognise that tension.

The Bank of Canada prints its dollars on paper plastic. The Bank of Montreal prints its dollars on silicon(?). But both print dollars. BoC dollars and BMO dollars have a fixed exchange rate of one-to-one.

What is it that makes the BoC, and not BMO, the one that chooses Canadian monetary policy? How come it is the BoC that chooses and implements the 2% inflation target, and not BMO choosing and implementing whatever target it is that BMO would choose?

This is my answer: Because it is BMO that makes the commitment to peg its dollar to the BoC dollar, and not the BoC that makes the commitment to peg its dollar to the BMO dollar. It's an asymmetric peg. Just like if the US Fed decided to peg the US dollar to the Canadian dollar, and the BoC made no such commitment in return, but did whatever it wanted to do, then US monetary policy would be set in Ottawa. That's what makes BoC the alpha bank and BMO the beta bank. The alpha bank chooses the direction for its money, and the beta bank just follows along.

But if there's a run on the BMO, and the BoC steps in to act as lender of last resort, their roles are temporarily reversed. It is now the BoC that does what is needed to ensure the exchange rate remains fixed at one-to-one.

If the BoC's commitment to act as lender of last resort to BMO were open-ended and unconditional, then BMO would become the alpha bank, and Canadian monetary policy would be set in Montreal Toronto, and not in Ottawa any more. BMO could do whatever it likes with its dollar, and the BoC would just follow along.

The only reason this does not happen is that the BoC's commitment to act as lender of last resort to BMO is not open-ended and unconditional. The BoC would insist that BMO get its act together soon, or shrink its balance sheet, or have its balance sheet forcibly shrunk.

It's the same if the BoC ever needed to act as lender of last resort to the Government of Canada, if there were a run on GoC bonds. If that commitment were open-ended and unconditional, and if the GoC took advantage of that commitment, the BoC would lose its alpha status. Another way to say the same thing is that the BoC would lose its independence.

Now imagine if the Canadian provinces took over the role of the federal government, leaving the BoC as the only federal institution. And imagine also that Canadian commercial banks were provincially based, with each province having its own separate set of commercial banks. Canada would look like the Eurozone, and the BoC would look like the ECB.

Now suppose that PEI's solvency was in doubt, and the solvency of PEI's banks was in doubt. Would the BoC make an open-ended and unconditional commitment to act as lender of last resort to PEI and its banks? If it did so, it might as well re-name itself "The Bank of PEI". Canadian monetary policy would be set in Charlottetown, not Ottawa. An independent BoC would not do this.

So I think I disagree with Simon Wren-Lewis on ECB independence.

On the other hand (there's always an other hand), there are times when acting as lender of last resort does not conflict with central bank independence. If Canadian inflation were at risk of falling below the Bank of Canada's 2% target, and if the government of Canada needed a lender of last resort to buy its bonds, then of course the BoC should go out and buy loads of GoC bonds. Which is exactly what it would do. And one can fairly criticise the ECB for not having done a lot more of something like this with a basket of Eurozone "provincial" government bonds.

Comments

[Avon Barksdale]

"But if there's a run on the BMO, and the BoC steps in to act as lender of last resort,...the BoC would insist that BMO get its act together soon, or shrink its balance sheet, or have its balance sheet forcibly shrunk."

And this is the central problem in finance today - the moral hazard the comes with a commitment to rescue banks. This is so harmful, so anti-market, so crony capitalist, that it is astounding that we can walk around this elephant in the room.

To tame the moral hazard we ask our central banks to become macro-prudential oversight bodies that work with regulators to manage bank assets. The problem is not the assets, loans or risk management, it's the source of funding - the bank's liabilities. Equity can't run, long term debt can't run. Get the banks to issue more capital!

We want a competitive, innovative banking system in which banks can fail without any systemic problems. Put the central bank on auto pilot - follow a Taylor Rule - and stop trying to engineering bailouts.

[Nick Rowe]

Avon: for once we basically agree. (But the Taylor Rule is not a very good autopilot).

Monetary policy (in the narrow sense) has a very transparent target, and the BoC can be held accountable for (more or less) hitting that target. And any smart high school student who knows no macro can spend a couple of hours on the BoC website and figure out the basics of how the BoC plans to hit its target. We have transparent rules (with a little bit of discretion in implementing those rules).

The lender of last resort function is the exact opposite. Not just moral hazard, but politicisation, and decisions made in secret in a great hurry, with no rules of the game.

[JKH]

In other words, the Bank of Canada operates within risk limits.

As does the ECB.

Independence does not mean either the option or the obligation to be reckless.

[Tom Brown]

Nick, somewhat O/T, but is this chart of inflation in Canada accurate? If so, why do you think it has dipped recently? I bring it up because you mention a 2% inflation target in your post (as you have many times in the past).

[Too Much Fed]

JKH, can we go over this?

Let’s assume BMO is the only commercial bank. Also assume a 0% reserve requirement and 0 desired reserves for BMO. BoC is an ECB type of central bank.

Next, assume the BoC can set the “fed funds rate” with an announcement affect, the BoC can set the penalty rate, and both of them are the same. BMO sets the rate for checking accounts and savings accounts at the same rate as the “fed funds rate”. BMO is always solvent.

Now have everyone want to use all currency on odd days and all demand deposits on even days. Currency and demand deposits have the same velocity. Demand deposits stay at 7 trillion, then become 7 trillion in currency, and then back.

The system runs for 30 days this way with no change.

After 30 days, entities start borrowing from BMO causing demand deposits to go to 10 trillion pushing price inflation above target. The next day all 10 trillion in demand deposits get converted to 10 trillion in currency, still above the price inflation target.

*** The way the system works now is BoC never threatens the 1 to 1 convertibility.***

BoC hopes market interest rates move up enough. If not, BoC could raise the penalty rate/”fed funds rate” high enough to invert the yield curve. I suppose it could raise the capital requirement. Let’s skip that for now.

“If the BoC's commitment to act as lender of last resort to BMO were open-ended and unconditional, then BMO would become the alpha bank, and Canadian monetary policy would be set in Montreal Toronto, and not in Ottawa any more. BMO could do whatever it likes with its dollar, and the BoC would just follow along.”

I say that is how the system works now. The only “conditions” are solvency and the penalty interest rate/”fed funds rate”.

“The only reason this does not happen is that the BoC's commitment to act as lender of last resort to BMO is not open-ended and unconditional. The BoC would insist that BMO get its act together soon, or shrink its balance sheet, or have its balance sheet forcibly shrunk.”

I am not exactly sure how BoC would forcefully shrink BMO’s balance sheet, but I am going to say threatening 1 to 1 convertibility is not on the list.

Thoughts?

[Andy Harless]

The thing about a basket of bonds is that you get more bang for the buck from some bonds in the basket than others. If you're trying to get the EZ inflation rate up to 2%, buying a German bond hardly makes a difference. Buying a Greek bond would make a much bigger difference. And in general, what with the zone as a whole being below target, anything you can do to help Greece is going to have a disproportionately large effect on the inflation rate for the zone as a whole (assuming Greece stays in the euro).

[ThomasH]

Right as far as it goes, but if and when BoC needed to bail out BoM's creditors (to the extent that they were not complicit in the problem being resolved) and re-organize the bank, it would do so in a way that did not destroy value. If the profits of the reorganized BoM continued to fall year after year following the reorganization, one would conclude that BoC had messed up the reorganization. And any subsequent reorganization -- which would probably require writing off some of the loans BoC made as part of the failed reorganization -- would look a lot different from whatever was done the first time around.

ECB's error in Greece -- over and above it's errors in not maintaining Eurozone NGDP growth since 2008 -- was to get involved in the dispute between the Greek government and its creditors when it's job was being a lender of last resort to the Greek banks. And since those creditors were largely private, the role of the EC in that dispute was doubtful as well except that it was their banking regulations that told creditors that loans to Greece were no more risky than loans to Germany.

[john]

"If the BoC's commitment to act as lender of last resort to BMO were open-ended and unconditional, then BMO would become the alpha bank, and Canadian monetary policy would be set in Montreal Toronto, and not in Ottawa any more. BMO could do whatever it likes with its dollar, and the BoC would just follow along."

This does not seem right.

If there is a run on BMO and BoC supplies liquidity, BoC obtains an option on the (new) equity of BMO. If liquidity needs are not worrisome, perhaps BoC will not exercise. But in your scenario, BoC would certainly exercise, in which case BMO becomes a branch of BoC.

In the EU (golly, these guys' politics is complicated!), ECB obtains no option. And is there no BK process to cram down creditors? A national bank cannot restructure its own country's banks? Debt is forever?

But for Canada, it seems BMO would toe the line in short order.

[Ralph Musgrave]

Nick,

You claim it’s the fact of the BMO making “the commitment to peg its dollar to the BoC dollar” that makes the BOC “the one that choses Canadian monetary policy”. A couple of other factors are relevant.

First, the Canadian government has declared that money issued by its bank shall be legal tender. In contrast, the BMO does not have that power. Second, the Canadian state has declared that taxes can only be paid using the state’s money.

[Brad Kitson]

Nick,
Simon is making a practical argument and you are making an abstract argument. It's non-sense to disagree with Simon on that basis.

Furthermore, Simon never has never argued that the ECB is an "open ended and unconditional" lender of last resort (lolr). Therefore, by your own logic, your condition for disagreement does not exist.

[Lord]

It depends on the relationship between the two. If the BoC exercises regulatory/supervisory control over BMO, there is none, or at least, needn't be any conflict. It is only if BMO operates entirely independently that their could be. A branch bank can get the parent in trouble, but the parent would be as much at fault for its failure.

[Nick Rowe]

JKH: the "risk" in question is the risk of not carrying out your own monetary policy.

Andy: you have a point, but "bang per buck" (or per Euro) is maybe not the best metric. Bucks are not a scarce good, if you can print them.

Brad: on rereading Simon, you may have a point. But I think the ECB's risk of becoming the beta bank if it kept on acting as LOLR to Greek banks went up considerably after the Greek government announced the referendum, letting the current agreement run out of time. If it hadn't capped then, with no limit in sight, the ECB effectively becomes a branch of the Bank(s) of Greece.

[Lord]

The ECB wasn't really created with either regulatory/supervisory control or lolr powers so I can understand their reluctance, but whether they should have those powers, whether they would have acted any differently, whether they should be acting differently now, are reasonable questions to which I would answer, yes, no, and yes if they had or want those powers, or no if they thought immediate default was a better solution than extend and pretend, which I lean towards, and that they may be more successful in obtaining those powers that way. It isn't apparent they even want those powers though, which I consider rather irresponsible but others may not. Anyway, 'will do everything' has been shown false.

[Rick B]

In a functional system, when Greek (private) banks got into trouble making (private) debt payments, the Greek Central Bank or the Greek Branch of the ECB (which is what is missing from your example) should have had the option to recapitalize and restructure their liabilities, or let them fail with an orderly wind down ... Regardless, the debt holders should not have been made whole. The continued demands that they are made whole, leveraged by the ECB withholding liquidity is an illustration that the ECBs lacks political (and as a direct result then, monetary) independence.

There is no scenario in which the Greek government should have been on the hook for the (private) bailout without taking an equity stake in the restrucured banks, just as there is no scenario in which the Greek government should have taken an equity stake unless the debts were restructured. The ECB should not have viewed the Greek Branch of the ECB as its own entity tied to the Greek State, because the not only is the Greek branch not an arm of the Greek State, it had no control over the mechanics of how the (private) recapitalization happened. The ECB should not be taking direction on monetary policy from the European Commission, nor should it make providing liquidity to Greek Banks be contingent on political reform. Under those conditions, how can anyone make the claim that the ECB is independent.

Imagine if the FED refused to provide liquidity to US banks unless OBAMA scaled back stimulus in 2008... It's unimaginable.

So the Greek arm of the ECB is wholly dependent on the ECB which is wholly dependent on the politics of the European Comission... This was doomed from the start.

[Nick Rowe]

Rick B: "Imagine if the FED refused to provide liquidity to US banks unless OBAMA scaled back stimulus in 2008... It's unimaginable."

Suppose the US government were defaulting on its bonds, and US banks held lots of US government bonds.

Take the admittedly extreme case: what would a truly independent governor of the central bank of Zimbabwe have done?

[Majromax]

I'm not quite convinced that an open-ended lending facility is always at odds with an independent, inflation-targeting central bank.

When there is a run on *all* Greek banks, doesn't this mean that there is an excess demand for the medium of exchange? If so, then shouldn't the ECB be providing liquidity not just as a last-resort, but also to fill this excess demand for the MoE?

If the excess demand is not Euro-wide, then we don't have the complete picture. Are Greek depositors withdrawing their Euros to deposit at foreign banks? If so, wouldn't those other nations have an excess supply of Euros, meaning the ECB would be in an ideal place to balance this?

(That said, the only reason this question makes sense is because the balance sheets of Greek banks don't look like the balance sheets of French or German banks. It's all well and good that "the ECB must never monetize debt", but that only works if all member banks down to the retail bank hold Euro bonds in the same ratio. Otherwise, even if only on the infinitesimal margin, emergency liquidity provided to one country is more of a subsidy than that provided to another.)

[Rick B]

Nick, a better analogy would be: What if South Dakota was defaulting on its dollar denominated state bonds, but in this analogy the reserve bank of South Dakota was considered an arm of the state government; and, had such an excessive amount of bonds because the FED bailed out privately run Chase South Dakota through the reserve bank. What would the central bankers at the Fed do? Would they impose capital controls on South Dakota and lift FDIC guarantees on depositors to ensure all debts were paid in full? It's a weird thing to think about because it's an impossible situation.

And not just because the FED, or the federal reserve bank of SD, would never assume all of the liabilities of a defaulting private bank (unless a former chase CEO was now the FED chair...)

States have balance budget provisions to mitigate just how fiscally irresponsible they can be, which allows the FED to operate under the assumption that Zimbabwe can't happen in SD, I.e. Falls in GDP are linked to reductions in state budgets so the state can't operate a 'printing press'. So even if the problems in South Dakota were wholly the result of Public mismanagement, which they likely were not, the only thing that the FED would require is a balanced budget (including interest payments) provision in order to lend without worry. And then... what if there were a excess demand in the rest of the U.S. for FED backed bonds... We'd be talking a primary surplus of less than 1% to pay off principal plus interest... That would be the thought process of an independent central bank.

Greece could only dream of that relationship right now.

[Nick Rowe]

Majro: "When there is a run on *all* Greek banks, doesn't this mean that there is an excess demand for the medium of exchange?"

It means there is an excess demand for one medium of exchange (currency) and an excess supply of another medium of exchange (chequable bank deposits). But it does mean the total supply of media of exchange will fall, so the central bank may need to take offsetting action. But it can't have an open-ended unconditional commitment to expand the supply of currency in proportion to the supply of deposits, without handing over control of monetary policy to the banks.

Rick B: yep, I think that works. It's like my PEI (a Canadian province) analogy, except you have a reserve bank of South Dakota in yours.

[Jim Rootham]

Just to be pedantic and silly about your silicon(?). BMO dollars are magnetic circles on rust (iron oxide) most of the time. Most of the rest of the time they are electrons on aluminum. They only live as electrons in silicon while they are being changed. All this assuming magnetic disk platters and MOSFET (CMOS) computers.

[Majromax]

> But it can't have an open-ended unconditional commitment to expand the supply of currency in proportion to the supply of deposits, without handing over control of monetary policy to the banks.

It can through the discount window, but only if it can make credible, forward-looking statements about what can be used for collateral.

In the case of Greece, a hypothetical such ECB would recapitalize the banks, but then it would announce that no Greek debt could be used as collateral until (blah blah). The Greek banks would be fine, but the Greek government would hang.

(Why do Greek banks hold predominantly Greek national debt, anyway? The other banks of the EU rebalanced their portfolios to avoid this in 2009-2012 or so. Is there some legal requirement that Greece's debt be held by its national banks? Does the Greek National Bank only accept Greek bonds for its own collateral? Why is a Greek bank different in operation than a German bank, aside from its lender of last resort?)

> All this assuming magnetic disk platters and MOSFET (CMOS) computers.

Super pedantic; aren't big database solutions like those of banks moving to SSD-backing for performance reasons? If so, those are bits-in-MOSFET, which would be electrons-on-silicon again.

[Oliver]

This is my answer: Because it is BMO that makes the commitment to peg its dollar to the BoC dollar, and not the BoC that makes the commitment to peg its dollar to the BMO dollar. It's an asymmetric peg.

Even if one assumes that inflation targets can always be implemented, your claim is actually stronger. Not only must inflation overall be 2% but each member bank (member country in the EZ) must live up this target individually.

Firstly, in the Eurozone that was never the case. Germany consistently undershot whereas the peripheral nations overshot in terms of inflation. The focus was on the average while differences were thought to take care of themselves. But still, there was always only one Euro.

Within a country the assumption that all banks are producing a homongenous currency substitute (or else...) might be somewhat admissable in that different banks are effectively competing for shares of the same market. Within the Eurozone that isn't really helpful, though. AFAIK, each bank can only compete for market shares within its own country of residence. And we all know there are small, large, rich, poor, industrialised, agrarian etc. nations within the EZ. Equality of outcomes is virtually impossible.

Secondly, and for the reasons above, I think an asymmetric peg is the wrong way to look at the division of labour between banks and their central bank.

I think it's OK to say that banks commit to achieving a common goal. But in the end, whether they manage or not, the CAD /Euro is the average of what member banks / nations are effectively issuing. There is no actual alpha currency there which to peg to. The value of the Euro is the average of all liabilities issued under its name. CB currency is actually only the commitment by the CB to keep up the illusion of parity between its issuing agents. The bigger the difference between its members, the bigger that commitment must be.

The Euro is actually very heterogenous for reasons that lie far beyond the mandate of the ECB. And, barring large fiscal transfers and something like Eurobonds, the only way for the ECB to paint over that inherent discrepancy is it to lend to its member NBCs open-endedly and unconditionally, I would say.

[JP Koning]

"then BMO would become the alpha bank, and Canadian monetary policy would be set in Montreal Toronto, and not in Ottawa any more."

So by raising interest rates, BMO could cause the entire Canadian price level to rise, and vice versa?

[Nick Rowe]

JP: "So by raising interest rates, BMO could cause the entire Canadian price level to rise, and vice versa?"

That should be "fall", not "rise". But yes. Because all other banks (including the BoC) would have to raise theirs, to prevent their money depreciating against the BMO dollar.

[Too Much Fed]

JP said: "So by raising interest rates, BMO could cause the entire Canadian price level to rise, and vice versa?"

Can you explain that in a little more detail?

One thing I am looking for is what interest rates are you referring to, loan rates or deposit rates (savings accounts / CD's).

[JP Koning]

Whoops, yes -- I meant to say fall. Ok, got it. Good post.

Too Much Fed: I was thinking the deposit rate.

[Jim Rootham]

To crank the pedantic (twice). Given reliability issues the ground truth is probably on platters, buffered onto SSDs, where the electrons are on aluminum (M for metal in MOSFET).

We are getting WAY over the top here.

[Nick Rowe]

Jim: "We are getting WAY over the top here."

Totally. But it is fun, in a weird way.

[Jacques René Giguère]

Nick: there is no wasdted knowledge. ANf given the madness around us, quantum physics is a haven of peace and rationality. Even the peace of commenting from the cottage terrace, looking at the sailboats on the St-Lawrence is not enough these days...

[Nick Rowe]

Jacques Rene: I know what you mean. Yesterday I went to Kenny U pull, and salvaged an O2 sensor off a junked car for $4, and fixed the check engine light on a 99 Tercel. Small victories help keep us sane.

Someone should trade his 1300cc Yamaha for a moped, to show a little humility.

[Too Much Fed]

"If the BoC's commitment to act as lender of last resort to BMO were open-ended and unconditional, then BMO would become the alpha bank, and Canadian monetary policy would be set in Montreal Toronto, and not in Ottawa any more. BMO could do whatever it likes with its dollar, and the BoC would just follow along.

The only reason this does not happen is that the BoC's commitment to act as lender of last resort to BMO is not open-ended and unconditional. The BoC would insist that BMO get its act together soon, or shrink its balance sheet, or have its balance sheet forcibly shrunk."

I was hoping JKH would comment, but I am going to try myself.

I am going to say the only condition is solvency of BMO. If BMO is solvent, then the lender of last resort is open-ended and unconditional. That means if there is a BMO run (demand for demand deposits falls and the demand for currency rises), the BoC will provide an elastic currency (increase the supply of currency to meet the demand) so the fixed exchange rate of 1 to 1 will be maintained even if the price inflation target is missed.

I would say “monetary policy” is carried out by BMO.

I think this is why Nick and I disagree so much (I say BMO demand deposits are both MOE AND MOA). However, I believe there are plenty of other people when you “quiz” them enough believe the same thing. Most of them are the “accounting” and “finance” people.

[Too Much Fed]

"JP: "So by raising interest rates, BMO could cause the entire Canadian price level to rise, and vice versa?"

That should be "fall", not "rise". But yes. Because all other banks (including the BoC) would have to raise theirs, to prevent their money depreciating against the BMO dollar."

And, "Too Much Fed: I was thinking the deposit rate."

First, I seem to remember reading somewhere that before the fed was created demand deposits of different commercial banks did not trade at a fixed exchange rate. After the fed was created, all member commercial bank demand deposits traded at a fixed exchange rate of 1 to 1 between each other both ways. With that in mind and along with believing that is the way the system works now…

Let’s see if I am getting this correct. Let’s say there is TD Bank, BMO, and BoC. BoC “fed funds rate” and “penalty rate” are both 3%. TD and BMO both pay 3% on checking accounts. BMO raises the checking account rate to 4%, while TD and BoC keep their rates the same. Lending rates of BMO and TD stay the same. Assume this causes entities to save more (it may or may not) and price inflation to fall.

Every entity transfers their checking account to BMO. The demand deposits “move” (as in mark down then mark up). The central bank reserves actually move. This happens at the fixed exchange rate of 1 to 1. Prices fall in terms of BMO, TD, and BoC currency. TD is now short of central bank reserves. It borrows BMO’s extra central bank reserves for 3%.

I believe that all that happens is price inflation can fall. BMO is less profitable. The checking account holders at BMO are more profitable.

Since the creation of the central bank, currency and all member bank demand deposits trade at a fixed exchange rate of 1 to 1 both ways. All member bank demand deposits trade at a fixed exchange rate of 1 to 1 between each other both ways.

The only thing the member banks try to do is remain solvent. There is no trying to maintain the exchange rate between currency and other demand deposits of member banks by some means. It is not like a foreign currency peg that is one way.

[Sina Motamedi]

Great post, Nick.

The Federal Government is the ultimate Alpha -- it could change BoC policy overnight if it REALLY wanted to. However, the Fed outsources policy to the BoC and uses "independence" to make sure long-term incentives are aligned. The Feds have learned to do this over time, and it is a very good thing.

Now, what would be the allocation factor for the basket of goods you speak of -- population, GDP? Wouldn't this, if you dig deep enough, simply be what we call "monetary policy"? And thus, that the euro crisis all boils down to bad ECB monetary policy.

[Too Much Fed]

Basically, an elastic currency (maintain a fixed exchange rate of 1 to 1 both ways between currency and demand deposits of the commercial banks) has priority over the price inflation target.

If I am remembering correctly, an elastic currency was part of the Federal Reserve Act in 1913, while price inflation targeting was introduced in the 1970's.

[Nick Rowe]

Sina: thanks!

I think the Eurozone would be in a lot better shape if the ECB had done better monetary policy, including (but not only) acting as a better lender of last resort. But the Eurozone is inherently flawed, and it is especially flawed when it comes to acting as lender of last resort, simply because it can't force the betas to get their acts together, so risks becoming a beta to their alpha.

[asp]

You're talking about Canadian Credit Unions, right?