« The degeneracy of FTPL | Main | Who would ever lend to an FTPL government? »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

When the federal government's Treasury is empty, and it has to pay back its bonds, it can force the bank to print money. In that event, the government is alpha, and the bank is beta.

But what if instead of a transparent and fully credible inflation target, the Alpha is kind of vague about its intentions? And if every so often its leadership gets replaced by someone appointed by the Beta? That's kind of how it seems in the USA. Might not the Fiscal Theory of the Price Level apply?

This is really a devastating (and very clear) critique of the FTPL. I also think it is fairly empirically certain that monetary base (currency and reserves) controls interest rates on bonds:

http://informationtransfereconomics.blogspot.com/2014/11/quantitative-easing-cleanest-experiment.html

As long as for political reasons , or because it thinks it is the right thing to do, (perhaps it reads this blog) the government allows an independent CB to set monetary policy then money is alpha.

However, at any time for political reasons or because it is driven by its own economic theories, a government can run whatever deficits it likes and mandate that the CB accommodate any money printing required to make these deficits sustainable. It seems to me that fiscal policy is alpha here.

I'm not sure the second case "proves" FTPL as at any moment the government could revert to the first case - but it does muddy the waters quite a bit.

This is a very clear exposition, but it actually made me raise my degree of belief in FTPL. Why would the central bank ever *not* bail out a government risking default? Is the central bank run by sociopaths who hate their country *and* also command an army more powerful than the government's? And if such a powerful junta did control the central bank, wouldn't it just overthrow the government and thus control fiscal policy as well?

First thoughts dump:

As you know Nick, I’ve never understand the asymmetric redeemability argument in the sense that I don’t think the language or the idea of redeemability in one direction but not the other is the way to express the pricing idea. The fact that a number of commercial banks carry the currency of a central bank as inventory does not mean that redemption is not symmetric. If I demand currency for my BMO deposit, I get it. If I demand a BMO deposit for my currency, I get it. The central bank makes sure that happens using standard operating procedure involving the commercial banks as satellite issuers of money with the same value. It works in both directions. When the Bank of Canada targets the value of the dollar, it’s targeting both sides of the redemption transaction in both directions, in effect.

The exception is if BMO goes out of business. But that’s a separate issue from the value of a dollar unit for either the currency or a BMO bank deposit.

So the value of the dollar as a unit is a separate issue from the quantity of BMO dollars (or currency on my demand) that I am allocated, whether BMO is a going concern or is wound up. And even if BMO went bankrupt and I got paid off “50 cents on the dollar” (assume no deposit insurance), it’s not really “50 cents on the dollar” - it’s 50 per cent of dollars on deposit. But each of those dollars still has the same value.

Without deposit insurance, there is no fiscal guarantee for the quantity of commercial bank dollars, but there is a monetary value guarantee for each dollar that remains.

With FTPL, there is a fiscal guarantee for all dollars, provided that the fiscal issuer takes a sensible view of having a “currency issuer” at its ultimate control, and doesn’t default on its liabilities. (Greece is not a currency issue in this sense. Canada and the US et al are.) And the fact that there is a fiscal guarantee probably has something to do with the value of the currency.

The government of Canada is not beholden to the Bank of Canada in this sense.

I don’t know whether any of this supports FTPL, but I don’t see it weakening the case.

Nick:

Here's a bank's T account

ASSETS.....................LIABILITIES
100 oz silver..............$100 in paper money
bonds worth $200...........$200 in paper money.

Let's say a dollar is pegged to 1 oz of silver. Then the $200 in the second line is issued, not for 200 oz worth of government bonds, but for $200 worth of government bonds. If the issuer of those bonds goes broke, and the $200 worth of bonds falls in value to $170, then let E=the exchange value of a dollar (oz./$). Setting assets (100 oz plus $170 of bonds worth E oz. each) equal to liabilities ($300 worth E oz. each) yields

100+170E=300E, or E=.77 oz/$

The bank can no longer afford the peg of $1=1 oz., and trying to maintain that 1:1 peg would only invite a bank run. The 10% loss of assets caused a 23% inflation.

So the government's fiscal health determines the value of the dollar, just like the FTPL says. Its the assets backing the dollar that matter, not the peg, and not the denomination of the assets.

IF the U.S. pegged it's currency to Canadian. Yes, it could happen.

More realistically, both countries are acting independently and competitively. We could even consider EACH as a separate COMPANY.

If each WAS a separate company, then the board of directors (the Central Bank) of each can decide on how to price products sold to every other company (other national economy). (The CB can do this by issuing more money as Japan is doing.)

No private company can control prices to the degree I have suggested. Central Banks have a tool not available to private companies, the tool of issuing money, which DOES give the CB's that ability.

Another subject: In my view, the method of expanding the money supply is critical to any discussion of money. I believe that bank lending expands the money supply only because more people think they have title to the same underlying money (which you might call BASE MONEY). I believe government borrowing has the identical effect.

Now to the distinction between bonds and currency. I just indicated a belief in both governments and banks as having the ability to expand the money supply. What happens to this money? Two choices:

1. The money disappears. No, the money disappears only when bank loans are repaid or government pays down it's bonds.

2. The money has NOT disappeared. Yes, the money must transform to look like neither a bank deposit or currency. The obvious possibility remaining is government bonds. Government bonds can be 'good as gold' and may even pay interest. They do have the problem of reduced liquidity but only reduced liquidity, not NO liquidity.

If we make the two basic assumptions that I propose here, the entire discussion becomes much different.

When you say asymmetric redeemability, I see you as just using another phrase to describe institutional behavior. Redeemability is fragile. Pegs can be broken. What matters is the perception the contingent behavior of these institutions regardless of current arrangements. Not to say they are irrelevant, but the direction of redeemability isn't guaranteed. That's the whole point of who acts last. Redeemability existed in both Russia and in Hungary. But that didn't make it clear who was going to be the alpha when fiscal situation got dire -- and the results were not the same. If people started thinking that the Bank of Canada cared more about the United Way than price level stability, the United Way would absolutely control monetary policy.

What about the zero lower bound? Recent history seems to demonstrate that the central bank can put a cap on inflation, but not necessarily a floor. For that you need fiscal policy.

Fiscal Aspects of Central Bank Independence

http://sims.princeton.edu/yftp/Munich/CBInd

All: please read my replies to dlr, JKH, and Arnold, together.

dlr: in the olden days, the dollar was pegged to gold. Gold was alpha and the dollar was beta. Then the dollar broke the peg, and became an alpha in its own right. Because the prices of goods were measured in dollars, not in gold. Yes.

JKH: if BMO breaks the peg, because it is unable or unwilling to redeem BMO dollars for BoC dollars, that means BMO is no longer beta. BMO dollars trade at a discount to BoC dollars. The beta wolf cannot keep up with the alpha wolf. But does that mean that BMO dollars become an alpha in their own right? No. Because (presumably) we would measure prices in BoC dollars and not in BMO dollars. BMO is not an alpha; it is a lone wolf, like Bitcoin, if it survives at all.

Arnold: Yes. If the beta wold will not or cannot keep up with the alpha wolf, the beta wolf might insist that the alpha wolf stay with the beta wolf. In this case the alpha wolf is no longer a true alpha. This is what happened in Zimbabwe. But most countries are not like Zimbabwe.

Mike: by far the biggest asset of the Bank of Canada is not the value of the bonds, or forex or gold or silver, (though technically the Department of Finance owns those), but the willingness of people to hold BoC currency at a negative real rate of interest (or a rate of interest less than the growth rate in the demand for currency). It can borrow against that asset, if it needs to.

Nick,

I have no problem with the alpha / beta language or the connotation.

It’s the asymmetric redeemability language and meaning that bothers me.

Alpha is the monetary policy bank. That says it all for me.

Beta is in effect an agent of alpha, distributing alpha money and issuing beta deposits that are denominated in beta money, where beta money and alpha money are exchangeable 1:1 per unit.

I.e., equivalence in the unit of account.

There is credit risk on beta deposits – unless fully insured (which tends to be the case).

Credit risk on a beta deposit is not the same thing as credit risk (or monetary risk) on a unit of beta money.

If there is a 50 per cent write-down of beta deposits, the 50 per cent that remains is still redeemable 1:1 for alpha money.

Both before and after that 50 per cent write-down, there is symmetric redeemability of the alpha monetary unit and the beta monetary unit. It’s the size of the beta deposit that has changed after the write-down – not the integrity of the exchange value per unit.

There’s a difference between monetary risk and private sector credit risk.

Just as there is a difference between monetary risk and fiscal risk.

Great pair of posts, thanks!

@Arnold "When the federal government's Treasury is empty, and it has to pay back its bonds, it can force the bank to print money. In that event, the government is alpha, and the bank is beta."

I second this (also after reading Nick's answer). It doesn't make sense to separate the central bank and call it alpha. E.g. in the US the FED derives its powers from the congress. So the congress/government would, being true alpha, just order and not "insist".

It is funny but Zimbabwe wasn't democracy. In democracy the central bank is strictly beta.

It is true that most countries are not like Zimbabwe. But the question isn't really whether countries are like Zimbabwe unconditionally. It is rather a conditional one: Can the central bank resist becoming like Zimbabwe's if the fiscal authority decides to become like Zimbabwe's?

@Dan:

> This is a very clear exposition, but it actually made me raise my degree of belief in FTPL. Why would the central bank ever *not* bail out a government risking default? Is the central bank run by sociopaths who hate their country *and* also command an army more powerful than the government's?

It didn't happen in Canada, in the 1990s. In that situation, the GoC was facing the spectre of a credit-rating downgrade driven in part by a large Federal deficit. Rather than mandate monetary financing, the government entered a period of severe retrenchment to reach a balanced budget position.

It isn't happening in the Euro area, right now. National bonds are priced with credit risk, on the assumption that the ECB won't bail them out, at least not without sterilizing the issue in some way.

It didn't happen in the US, during the last debt crisis. There, the country flirted with actually defaulting rather than issue further debt, which is still several steps removed from monetary financing.

The critique of "the CB will follow the government" is that it's looking at only one side of the conditional probability: the probability that the CB will let the government default provided it acts irresponsibly. This ignores the probability of the government acting irresponsibly in the first place, and here history has shown that when push comes to shove many governments retrench.

This is problematic from the FTPL standpoint, because this means that the demanded flow of seniorage revenue is not well-constrained in the future. If there is a 95% probability of the amount being finite in a nominal sense and a 5% probability of it being nominally infinite (hyperinflation), then the present value of future deficit is itself undefined.

@JKH:

> If I demand currency for my BMO deposit, I get it.

No you don't. You only get currency for your BMO deposit when you go to BMO and ask, or do so by proxy with an Interac transaction or cheque. The stable financial system means that intermediates (cheque recipients, Interac) are willing to be your proxy, but this is a deliberate construction of policy and not necessarily an economic certainty. This is aided by deposit insurance, but systems such as this also incorporate regulations to ensure that the potential liability is limited.

@Mike Sproul:

> Then the $200 in the second line is issued, not for 200 oz worth of government bonds, but for $200 worth of government bonds.

There's your problem. Government bonds on the books represent a limited commitment by the central bank to monetize debt in the event of default, since real-world central banks do not have much power to unilaterally and instantaneously withdraw money from circulation. (They can do so over time by withholding seniorage payments).

However, if we're assuming that a CB is acting as alpha to the government bonds' beta, then the amount of bonds on the CB's balance sheet is not a function of how much debt the government is issuing.

This interrupts the FTPL thinking, because only the portion of government debt that the CB chooses to hold can possibly be (is) monetized. Debt held by the general public is not monetary.

@JKH:

> If there is a 50 per cent write-down of beta deposits, the 50 per cent that remains is still redeemable 1:1 for alpha money.

And that means that the beta deposit is beta because the alpha money remains the unit of account.

If instead the beta money were the unit of account, then the nominal value of the deposit would not be written-down, but instead the price level (in alpha terms) would change.

In another perspective, "this beta deposit made on 1 January 1990" is not always redeemable at par, since its nominal value may change over time. "This beta deposit marked to market at the current value" is redeemable, but that's because we've already done the hard work of discounting.

Reading through the comments here, I think we need to be more precise about what the "federal government" means. There's the national legislature as a political entity (What is it called in Canada? Parliament?), and then there's the national treasury as a financial entity. The latter, which does the borrowing, is ultimately a slave to the former. And the central bank is also ultimately a slave to the legislature. When push comes to shove, the choice that the legislature would make among its two slaves is not obvious.

Suppose, for example, that in some future dystopian scenario, the US Treasury found itself unable to borrow and the Fed, facing a legitimate concern about hyperinflation, refused to purchase enough of its securities to rectify that situation. Would Congress then pass a law requiring the Fed to bail out the Treasury? If it's a choice between the credit of the Treasury and the credibility of the national currency, would Congress necessarily choose the former? That's far from obvious. But FTPL depends on its being a foregone conclusion.

I'm running in the Martian Marathon, and because it's Mars, I need an oxygen supply. The flow rate of oxygen is controlled remotely by a doctor who cares about my welfare above all. Before the race, we agreed that the rate of oxygen flow should not exceed a certain threshold X because it would be bad for my long-term health if it did. This in turn imposes a limit on how fast I can run. If I were to run faster than this limit, my doctor would have to increase the flow rate of oxygen beyond the safe threshold, or else I would drop dead on the spot. Much as I would like to win the race, I care more about my long-term health and choose never to exceed the speed limit.

Who is in control of the oxygen flow rate?

Nick:
"the biggest asset of the Bank of Canada is not the value of the bonds, or forex or gold or silver, (though technically the Department of Finance owns those), but the willingness of people to hold BoC currency at a negative real rate of interest"

The bonds are right there for everyone to see, but the "willingness to hold" is an imaginary concept. So why give all the weight to "willingness"?

The bank of Amsterdam (est. 1609) held 100% reserves of guilders on deposit, and charged 2% storage fees. People traded with guilder deposits. You could claim that the BoA's biggest asset was "willingness to hold", but you'd have a hard time explaining why the BoA bothered to hold any assets at all, or why the stock of guilder deposits was worth exactly the same as the guilders held in the vault, or why the BoA guilder lost value when, after a couple of centuries, it was discovered that Amsterdam's officials had embezzled funds.

Is there any important difference between your claim and a (Clearly mistaken) claim that GM's biggest asset is the public's willingness to hold GM stock?

There's also the problem of what happens if willingness to hold falls by 30% or so. The bank would then find that it needs to have those bonds in order to buy back the refluxing money.

@majromax

I agree with Danzyn. One of the basic duties of a central bank is to support government debt. Because of the potentially inflationary consequences of this duty, the CB and Treasury usually get together to figure out how to manage the government debt situation when there's a risk of inflation, but basically this is a completely cooperative effort and has been since the birth of central banking.

The examples you give are just examples of Treasury and Central Bank cooperating in their management of the tradeoffs between issue-credibility-inflation for fiscal policy.

I have difficulty envisioning the "pure" fiscal policy that Nick Rowe is positing. To me it's impossible to talk about fiscal policy without talking about central-bank supported fiscal policy. This doesn't make the central banks alpha -- it just means that when functional governments (not Zimbabwe) use fiscal policy, there is a significant degree of cooperation and coordination between the central bank and Treasury.

Majromax:
"Government bonds on the books represent a limited commitment by the central bank to monetize debt in the event of default,
since real-world central banks do not have much power to unilaterally and instantaneously withdraw money from circulation."

I don't get your first line. Whose default? The government's?

As for the second line, every time a central bank conducts an open market sale of bonds it instantly and unilaterally withdraws money from circulation.

> I have difficulty envisioning the "pure" fiscal policy that Nick Rowe is positing. To me it's impossible to talk about fiscal policy without talking about central-bank supported fiscal policy.

Why not? Isn't this exactly what's happening in the Eurozone, or in Canada at the level of provincial governments?

@majromax

The ECB has taken dramatic measures to support government debt, even though it has no duty (in theory) to do so, because of the national debt/EU bank distinction. The structure of the EU/ECB have well-understood weaknesses, and the ECB is doing its best to square the circle, by not leaving government debt in the lurch.

I don't know much about Canadian provincial governments, and its not clear to me how they could ever engage in fiscal policy (but maybe that's because I have U.S. government structures as my framework). I don't see any reason why a central bank would have a duty to support provincial debt -- which is why I don't see how provinces can engage in fiscal policy.

Jussi,

"E.g. in the US the FED derives its powers from the congress. So the congress/government would, being true alpha, just order and not "insist"."

(1) 'Fed' is an abbreviation, not an acronym, and so it does not need to be capitalised.

(2) Congress could change the institutional set-up. This is not what the FTPL is about, because no-one denies that possibility. Similarly, Nick Rowe could lead a revolution and force the Bank of Canada to buy his debts; this possibility does not prove the NRTPL.

Danyzn,

"Can the central bank resist becoming like Zimbabwe's if the fiscal authority decides to become like Zimbabwe's?"

Yes, in the same way that it can resist becoming like Zimbabwe's if Nick Rowe decides to become like the Zimbabwe fiscal authority.

Nick Rowe,

It almost seems that you may have well have not bolded the salient section of your post, because a lot of people just seem to be ignoring it entirely!

Just to check if I understand: let's consider an anarcho-capitalist country in which the Bank of Rothbard issues the unit of account, and buys/sells private assets. The private agents of this country are pegging their debts to the Bank of Rothbard's unit of account. The unit of account is a fiat currency, but the "fiat" in question is purely one of private agents choosing to use the Bank of Rothbard's unit of account to price their debts. It is POSSIBLE that, via coercion or persuasion, one or more private agents might influence the issuance policy of the Bank of Rothbard, but in the absence of such coercion or persuasion, it's the Bank of Rothbard that is determining the price level.

Then we can add a state to the story, which also issues debts denominated in the Bank of Rothbard's fiat currency. The price level is still determined by the Bank of Rothbard. It is POSSIBLE that, via coercion or persuasion, the state might influence the issuance policy of the Bank of Rothbard, but in the absence of such coercion or persuasion, it's the Bank of Rothbard that is determining the price level.

Am I on the right track?

"To me it's impossible to talk about fiscal policy without talking about central-bank supported fiscal policy."

It's possible. Just think of those writing about fiscal policy before the existence of central banks, or under a metallic standard, or all that has been written about constant-money supply fiscal multipliers.

W Peden: you are totally on the right track.

Though the Bank of Rothbard might also succumb to "Samaritan's dilemma" (pdf), as well as coercion or persuasion.

But which bit do you think I should have bolded?

Sorry, I said "may have well" when I meant "may as well". I was referring to the-

"The only thing that matters is whether those who issue financial assets peg the exchange rate of their financial assets to the financial asset issued by the Bank of Canada, and not vice versa. That's what makes Bank of Canada currency the alpha, and every other financial asset the beta."

- bit, which has been rather overlooked in the discussion so far, despite your bolding it!

@W. Peden

I am certainly willing to grant, as a theoretic matter, that we can talk about "pure" fiscal policy. But as a practical matter, if we are talking about the use of fiscal policy in the U.S. or in Europe, how do we discuss it without discussing central-bank supported fiscal policy?

My question is whether it makes sense to assume that people who advocate fiscal policy as a solution are not simply assuming the modern norm of central-bank supported fiscal policy (or in other words of a cooperative effort between Treasury and the central bank to make use of fiscal policy).

Perhaps the problem is that I am unfamiliar with the FTPL literature that this post is designed to rebut -- and in this sense Nick Rowe's point is much narrower than the question that appears to me to be interesting.

Mike,

"As for the second line, every time a central bank conducts an open market sale of bonds it instantly and unilaterally withdraws money from circulation."

I completely agree. I would add that the money has not disappeared, it has been bought into storage.

I think the correct way to look at this event is to say that the government does not want you to spend your money. Instead, the government wants to spend it right now and will store your money (in the form of a bond) until some future date.

This series of events should be considered as expanding the money supply because:

1. The bond purchaser knows his money is safe and will be available with only a very short delay.

2. The base money is spent by the government into the hands of more people who (in turn) may loan to government in future bond sales.

These blog discussions need consistent base assumptions. The assumption that money DOES DISAPPEAR when it is loaned to government is inconsistent with the bank loan concept that banks CAN EXPAND the money supply. After all, if a bank lends to government or an individual lends to government, either way, government gets money to spend.

Both bank and individual think that the bond received in exchange has enough value and security to count as the securest of reserves. The government bond is a good as cash except for a very short liquidity delay. Government bonds need to be counted as money supply.

"Perhaps the problem is that I am unfamiliar with the FTPL literature that this post is designed to rebut"

Maybe.

Nick,

I just re-read your post as I think I missed the point the first time round (I think you lulled me into a false sense of security with "If you understand the title you understand the post. It's obvious really.")

Is your point something like "even in cases where the people implementing monetary policy are consciously trying to assist fiscal policy goals they are, by adjusting the size of the monetary base, still implementing monetary policy and as such it is this (and not the fiscal policy goals they strive to help be met) , that are ultimately driving the price level (and people's expectations about the price level). Therefore FTPL is wrong" ?

Nick,

I just re-read your post as I think I missed the point the first time round (I think you lulled me into a false sense of security with "If you understand the title you understand the post. It's obvious really.")

Is your point something like "even in cases where the people implementing monetary policy are consciously trying to accomodate fiscal policy goals they are, by adjusting the size of the monetary base, still implementing monetary policy and as such it is this (and not the fiscal policy goals they strive to help be met) , that are driving changes in the price level (and people's expectations about these changes)" ?

“The only thing that matters is whether those who issue financial assets peg the exchange rate of their financial assets to the financial asset issued by the Bank of Canada, and not vice versa. That's what makes Bank of Canada currency the alpha, and every other financial asset the beta.”

Yes.

And that’s why "asymmetric redeemability" is a mischaracterization and red herring.

Alpha is principal and beta is agent.

The principal agent relationship is asymmetric in responsibilities.

Within that asymmetric arrangement, beta is the agent for symmetric redeemability. That is essential for the preservation of value. It is a crude case of open market operations – delegated to the agent.

The derivation of value is asymmetric, but it requires redemption symmetry in the implementation.

The important redemption symmetry occurs at the level of the beta agent (beta money for alpha money and vice versa).

There is a minor redemption symmetry at the alpha level (two different forms of alpha money, which is just an aspect of principal agent mechanics).

MF: I think that's a good reading of my post.

JKH: If I want to keep the value of Nick Rowe $20 IOUs fixed at $20 Bank of Canada notes, so my notes can neither fall below par nor rise above par, then yes I must redeem symmetrically. I must both buy and sell my IOUs at par on demand and supply. But it is me that makes that commitment. The Bank of Canada makes no such commitment to swap its notes for mine at par. That's the asymettry.

"2. The Bank of Montreal pegs the exchange rate of the Bank of Montreal dollar to the Bank of Canada dollar. Not the other way round. The Bank of Montreal promises to redeem its dollars at par with Bank of Canada dollars. Not the other way round. The Bank of Canada is alpha, and the Bank of Montreal is beta. It's the same for all the other commercial banks. It does not matter whether the commercial banks are bigger than the Bank of Canada. That's why Canadian monetary policy is set in Ottawa, not in Toronto."

I'm agreeing with JKH.

As long as BOM is solvent, it is "both ways". For example, say there is $1 trillion in BOC currency and $6 trillion in BOM demand deposits. Next, BOM demand deposits go to $13 trillion. Now have every entity prefer to have BOC currency. BOC currency goes to $14 trillion.

BOC does not say well that will lead to too much currency and price inflation so we (BOC) will not allow the redemption.

Nick's post said "JKH: If I want to keep the value of Nick Rowe $20 IOUs fixed at $20 Bank of Canada notes, so my notes can neither fall below par nor rise above par, then yes I must redeem symmetrically. I must both buy and sell my IOUs at par on demand and supply. But it is me that makes that commitment. The Bank of Canada makes no such commitment to swap its notes for mine at par. That's the asymettry."

I agree that BOC does not guarantee a fixed conversion rate for Nick Rowe demand deposits and/or Nick Rowe bonds. BOC also does not guarantee a fixed conversion rate for BOM bonds.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad