I am going to start with an orthodox monetarist approach, make one trivial semantic change, and see how far I can get in deriving Neo–Chartalist results. The semantic change is to change what I mean by "fiscal policy". It's an unconventional definition of fiscal policy, from a monetarist perspective, but I don't think a monetarist could otherwise object.
Let's start with the short run government budget constraint.
G-T +iB = dB + dM
Where G is government spending, T is taxes, i the interest rate on B, the existing stock of government bonds, dB the change in B, and dM the change in the stock of central bank money.
Fiscal deficits must be financed by issuing bonds, or money, or some combination of the two.
One conventional definition of "fiscal policy" is a change in G or T holding M constant (bond-financed deficits). A change in M counts as "monetary policy". Suppose instead we define "fiscal policy" as a change in G or T holding B constant (money-financed deficits). A change in B is what now counts as "monetary policy".
What the new dictionary calls "fiscal policy" the old dictionary considers as fiscal policy supplemented by monetary policy. And what the old dictionary calls "fiscal policy" the new dictionary considers to be fiscal policy with a countervailing monetary policy.
How would monetarism change if monetarists were forced to translate their beliefs into the language of the new dictionary?
Fiscal policy would have a very powerful effect on aggregate demand. (Sure, because it means increasing the money supply.)
Fiscal policy is unconstrained by the government's ability to borrow. (Of course it isn't, if it's financed by printing money rather than by borrowing.)
Providing it is not pushed too far and creates inflation, fiscal deficits impose no burden on future generations of tax-payers. (Sure, because deficits are financed by printing money rather than borrowing, so there's no future debt to pay; but watch out for the inflation tax if you do push it too far.)
By accepting goods (including labour) in lieu of taxes, at some fixed equivalence, fiscal policy can determine the price level. (Sure, because "taxes" mean a reduction in the money supply. It's just like the gold standard, where you can peg the price of gold by exchanging gold for money at a fixed rate, only it doesn't have to be gold of course.)
In general, once he realised that Neo-Chartalism uses "fiscal policy" to mean "money-financed fiscal policy", I don't see how a monetarist would really object to Neo-Chartalist policy recommendations. If we found ourselves well below "full-employment" (you mean below the natural rate), where there was no risk of inflation, the Neo-Chartalist's "fiscal deficits" (you mean helicopter money) would be supported by both.
Now, there are some Neo-Chartalist beliefs that may indeed be substantively different from (say) Monetarist beliefs. For example: those concerning the role of the state in determining what people use as money; and how paper money can have value. Those would need another post. But as for fiscal and monetary policy, Neo-Chartalism looks to me a lot like Monetarism plus a new dictionary. (OK, maybe add a Paleo-Keynesian reverse-L-shaped Aggregate Supply curve too.)
It would be equally possible, of course, to argue that Monetarists are really Neo-Chartalists plus an old dictionary.
I expect a Neo-Chartalist might object that fiscal policy really is money-financed, and this is not just some semantic question. And I wrote my post on Churches and Central Banks partly in anticipation of such a response. But suppose the currency printing press really really is located in the Treasury rather than in the central bank. That might matter in a game-theoretic analysis of the interaction between central bank and Treasury, if the two players have different objectives. But otherwise, so what? Why does it matter?
And where is the printing press located that prints bonds? Isn't that located in the Treasury too? Because if both are located in Treasury, how can the central bank do anything? It can't print money to buy bonds, and it can't print bonds to buy money.
Or perhaps both Treasury and central bank have a currency printing press? If so, should we not be looking at the game of chicken if the two players have different objectives? Because if they share the same objective, one press is redundant.
(I admit I'm new to Neo-Chartalism, and am almost certainly misunderstanding some of it. Several decades ago I did try to read Knapp's State Theory of Money, finding it not worth the candle, and actually did read and enjoyed reading Lerner's Functional Finance. This critique is part of my attempt to understand it better.)
Worthwhile Canadian Initiative 
Nick’s post said: “dM the change in the stock of central bank money.”
“Fiscal deficits must be financed by issuing bonds, or money, or some combination of the two.”
“Suppose instead we define “fiscal policy” as a change in G or T holding B constant (money-financed deficits). A change in B is what now counts as “monetary policy”.”
Can you define central bank money completely?
With my definition of money and with legal tender laws, I consider currency, currency denominated gov’t debt, and currency denominated private debt to be money (that is they are all mostly fungible).
Again, “A change in B is what now counts as “monetary policy”.”
So does that make “monetary policy” a change in currency along with a change in currency denominated private debt?
Hi Nick,
Firstly to understand this at a formal level requires you to read a bit of Wynne Godley and Marc Lavoie’s book Monetary Economics or you can check this post from Billy Blog Stock-flow consistent macro models Wynne and Marc won’t call themselves Chartalists (I presume) but FAPP their approach is close to it.
Your observations are true but I will not agree with your conclusions. When I started to read this, I too thought that it is Monetarism in disguise, though the only thing I know about Monetarism is Helicopter Money and they also know that much I believe.
Yes, fiscal deficits increase the money supply. But, banks loans do that too! If we really want to control the money supply we should control loans too. Please note – central banks do not and cannot control the money supply – remember they stopped publishing M3 ?
Printing money conveys a very emotive summary of the situation and it is better to not use the term. A better way to say it is that is
The central bank supplies as much currency notes as is demanded by households .
Btw, Godley-Lavoie’s books looks at this very closely and the changes in balance sheets of central banks, banks and the Treasury.
So let us first see (and hopefully agree) on how government spending happens. The Treasury has an overdraft at the central bank and atleast in the US will never bounce Treasury checks. If I work for the government, and when the payment time arrives, the central bank instructs my bank to increase my deposits by say $100. This is a liability for the bank and the bank needs to be compensated for this. The central bank, thus, increase the reserves by $100. The central bank will ask (and insist) the Treasury to raise some bonds in the market and have something in reverse – reserves go down by $100, deposits go down by $100 but the private sector has $100 worth of bonds. The sequence of events is simplified here and in practice, the Treasury is needed to issue bonds first and then spend. However the reserve position of commercial banks is so tight that the central bank provides the banks with some advances so that banks can buy the bonds themselves or don’t land into reserve requirement troubles when non-banks buy the bonds. Either way its the same.
This may appear trivial but is not so. Please note – everything happens electronically no printing takes place and as I have said earlier in the post,
The central bank supplies as much currency notes as is demanded by households .
Please note even James Tobin couldnt not get this when he worried about this. Others don’t even worry about it.
The attitude of Chartalists as far as debt is concerned is entirely different. Public debt is just a number even if it is a Japan like situation where the debt is huge and the inflation is little. The public debt is just a number!! There is no need to jack up prices and reduce the debt. In fact it is better to not jack up the prices. When payment time arrives, the treasury asks the central bank to just increase the reserves and deposits!
Here is what they say about “printing money”:
“Invoking the “evil-sounding” printing money terminology to describe this practice is thus very misleading – and probably deliberately so. All transactions between the Government sector (Treasury and Central Bank) and the non-government sector involve the creation and destruction of net financial assets denominated in the currency of issue. Typically, when the Government buys something from the Non-government sector they just credit a bank account somewhere – that is, numbers denoting the size of the transaction appear electronically in the banking system.
It is inappropriate to call this process – “printing money”. Commentators who use this nomenclature do so because they know it sounds bad! The orthodox (neo-liberal) economics approach uses the “printing money” term as equivalent to “inflationary expansion”. If they understood how the modern monetary system actually worked they would never be so crass.”
…and about money-financed deficits:
“Accordingly, the concept of debt monetisation is a non sequitur. Once the overnight rate target is set the central bank should only trade government securities if liquidity changes are required to support this target. Given the central bank cannot control the reserves then debt monetisation is strictly impossible. Imagine that the central bank traded government securities with the treasury, which then increased government spending. The excess reserves would force the central bank to sell the same amount of government securities to the private market or allow the overnight rate to fall to the support level. This is not monetisation but rather the central bank simply acting as broker in the context of the logic of the interest rate setting monetary policy.”
“…This also illustrates that government spending is independent of borrowing, with the latter best thought of as coming after spending.”
http://bilbo.economicoutlook.net/blog/?p=381
This is getting interesting!
First, on “printing money”:
Good finds himaginary! But contrary to Bill Mitchells’ (Billyblog’s) perhaps understandable fears, I am not using the “printing money” terminology to make it sound evil. I strongly support printing money in certain circumstances (like the last year or so). There is nothing wrong per se in printing money, only with printing “too much” money, and what counts as “too much” depends on the circumstances (will it cause inflation to get too high?). Fairly orthodox monetarist position I’m coming from there.
Ramanan: yes, some “printing money” really is printing paper currency, and some of it is electronic. I am using “printing money” metaphorically to cover both, because I think the physical form (paper or electrons) doesn’t matter for this question, and because I find it easier mentally to think of paper.
I haven’t read Wynne Godley and Marc Lavoie’s book. But by sheer chance I think I am nevertheless quite familiar with its approach! You see (presumably on Marc’s recommendation) I have been on the committee for several of Marc’s PhD students (long story), so I have seen the Godley/Lavoie approach at second hand, so to speak. Plus I have seen some of Marc’s seminars. My biggest problem with that approach, from what I have seen, is the typical assumption that the price level is exogenous, even in what I think of as the very long run (i.e. when stocks of assets etc have reached stationary equilibrium levels). That’s the Paleo-Keynesian bit of Post-Keynesianism/Neo-Chartalism, and I don’t like it.
Ramanan: what’s “FAPP” mean? Is it some technical economics term, or some texting shorthand?
himaginary: second quote from bilbo: I gave two definitions of “fiscal policy” (and “monetary policy”) in my post. I could have given many more. A third one I nearly gave is this: “Fiscal policy” means changes in G and/or T with M and B changing as required to keep i constant. Just a third dictionary, and one that Neo-Wicksellians like best. But Neo-Chartalists can’t keep switching dictionaries like that and expect us not to get confused about what they mean.
Returning to the question of who really really prints money (Ramanan’s “So let us first see (and hopefully agree) on how government spending happens.” paragraph). I think that what is at root here is the philosophical distinction between actively doing something vs letting something happen. I need to try to think up a good example to explain what I am trying to say more clearly.
Too much Fed: If you think M and B are equivalent, then there is only fiscal policy, defined as any change in G and/or T regardless of what happens to the composition of M+B.
By the way (slightly off-topic): I think it’s a bad rhetorical move to shy away from a term (like “printing money”) just because it has pejorative connotations. Much better to take the bull by the horns, embrace the term, and argue against those connotation. “Yes, we do advocate printing money; what’s wrong with that? No it won’t cause inflation, unless we print too much given the circumstances, except in the sense that it will prevent what would otherwise have been deflation”.
OK, I have thought of an example. Another automotive analogy, sorry.
In physical terms we observe a car starting to go downhill, the position of the gas pedal staying the same, and the car’s speed increasing.
But what theoretical account do we give of what happened? A cop issuing a speeding ticket would say the driver caused the car to speed up. A group of engineers testing a prototype would say the hill caused the car to speed up.
It all depends on what the observer thinks the driver ought to be doing. The cop thinks the driver ought to be keeping to the speed limit. The engineers think the driver ought to be keeping the gas pedal constant (that’s what they told him to do, so they could test the car).
If you think the central bank ought to be holding M (base) constant, the change in M is caused by the central bank. If you think the central bank ought to be holding B (or i) constant, any change in M with no change in B (or i) was caused by something else.
Our accounts of what happens are laden with a theoretical perspective of what ought to happen.
Two cars heading towards each other. Car A is on the left hand side of the road; car B is on the right hand side of the road. Which car caused the crash? Canadians say A; Brits say B.
Hi Nick . . . my first post here. I’ll say up front that I’m a card-carrying Chartalist or neo-Chartalist or whatever one wants to call it. I absolutely appreciate your very honest attempt at understanding and critiquing Chartalism here . . . as they say, at worst it’s still better than being ignored. I’ve heard that you were quite even-handed in your approach to these matters, and your piece here supports what I’v heard.
Ok, for my responses:
1. Very little of the discussion thus far resembles Chartalism, actually. The distinction isn’t b/n printing money or selling bonds, but rather between a deficit (however “financed”) and no deficit. That’s the objection to the term “printing money” . . . it applies to a gold standard or currency board monetary regime only. The federal govt ALWAYS spends by crediting reserve accounts of banks and recipients’ deposits. The decision to issue bonds is about supporting the interest rate, not “financing,” as without either the govt or the cb selling the bond (or otherwise paying interest on reserves) the overnight rate would fall to zero. Also, whether the govt sells bonds or “prints money,” the deficit has raised the net financial assets of the non-govt sector . . that’s just accounting 101. If anything, the deficit accompanied with bond sales is more “inflationary” than without, as the former adds an additional interest payment to the spending. More to be said here, but a few other good pieces are by Warren Mosler here:
http://www.moslereconomics.com/2009/02/19/deficit-spending-for-dummies/
http://www.epicoalition.org/docs/exchange_rate_policy_and_full_em.htm
2. The Godley/Lavoie models at their core are about getting the accounting right, not fixed prices. Every transaction in the economy affects the financial statements of those involved, and if you don’t get the effects on those financial statements right, then the subsequent analysis is inapplicable. One can insert any approach to price determination that one wishes; your critique seems rather tangential to me, therefore.
Ok, I’ll stop there for now.
Best wishes,
Scott
Hi Nick,
Nice discussion. FAPP = For All Practical Purposes.
Marc and Bill Mitchell are friends I think – learned it from one of his blog posts. Not suprising since the whole gang is centered around the Levy Institute (levy.org).
A bit about the Godley-Lavoie approach and the Post Keynesian and Chartalist approach in general. I couldn’t catch what you meant by price levels being exogenous. They have models with layers of complication. Prices are determined because of wage levels – the struggle between capitalists and workers. “When the capitalists and workers clash, its the price that suffers” – I made that up. Production firms raise money by bank loans to purchase fixed capital and pay wages in advance. They also accumulate inventories. Based on expected sales, target for profits, interest rate payments on loans, they do a markup pricing. This is very true of what we see in the world. The price of a toothpaste is not decided in some auction. The only place where supply-demand exists is in the financial markets. Now, more complicated models do not even have an equilibrium
The model deployed here describes a growing economy which does not spontaneously find a steady state even in the long run, but which requires active management of fiscal and monetary policy if full employment without inflation is to be achieved …
I know what you are trying to say. Tobin also said the same. A budget deficit is financed by a combination of printing money and selling bonds. This is not what Bill’s Blog says or prescribes. In economics, I have seen that nontautologies are branded tautology. For example if the government prints, who takes the money ? The central bank just prints enough to satisfy the daily cash requirements of the citizens. Government payments create both assets and liabilities at the same time. Assets for the holder of the deposits and liability in the form of reserves of the central bank. So it hardly needs to print. You may see printing happening in the data because households need it, not the other way round.
A bit about inflation. It happens not because of any printing money or deficits because of the wage struggle between capitalists and workers. It also happens because of energy prices and it can also happen because of supply side issues. In lot of cases (most?), the central bank doesn’t get this and just raises overnight rate targets.
A commentator in Bill’s blog (RSJ) said it nicely – the neolibral model is based on assumption of the zero sum game and the world is hardly a zero sum game. For generations we have been seeing an improvement in the quality of living. And how to make it a non-zero sum game – deficits!
One may ask, what if the government cannot raise debt ? This is hardly a problem – government spending automatically creates a liability for itself through spending and it does not really need to “borrow”. However, with the central bank targeting an overnight rate, it may need to. The solution is simple as well – pay interest on reserves. In other words, issue one day T-bills if you like it that way. Its good to think of the central bank and the government as one entity even if you dont like it.
Nick,
A few points. Like Scott Fulwiller, I am also an adherent to the Chartalist school of thought. But as Scott points out, what we’re discussing here is not some weird form of “high Keynesianism” but double entry bookkeeping, which has been in existence for several centuries. The second point is that Chartalists do not construct policy on the basis that a government faces NO CONSTRAINT in terms of spending, but that many of the “constraints” that people normally discuss are not identified in an honest manner – which is to say that these commentators import nonsensical notions of “affordability” and “national solvency” (we’re getting a lot of this sort of talk today in regard to the US and Japan), when they have no applicability in a post-gold standard world in which government alone creates currency and extrinsically confers value on it via the imposition of a tax liability. The public would not give up goods and services to the government in return for otherwise worthless coins or paper notes unless there were good reasons to do so. The primary reason the public accepts what we call “fiat money” is because it has tax liabilities to the government. If the tax system were removed, the government would eventually find that its fiat money would lose its ability to purchase goods and services on the market. In the words of Abba Lerner (one of the architects of “Chartalism”): “The modern state can make anything it chooses generally acceptable as money…It is true that a simple declaration that such and such is money will not do, even if backed by the most convincing constitutional evidence of the state’s absolute sovereignty. But if the state is willing to accept the proposed money in payment of taxes and other obligations to itself the trick is done.”
Once we accept the reality that the government creates new net financial assets exogenously, it renders notions of “solvency” to be nonsensical. It also means that the household analogy is totally fallacious, given that a household does not have the power to create currency or tax and therefore faces an external constraint unlike a government. Which brings us to the second point of chartalism. Again, as Lerner argues: “The central idea is that government fiscal policy, its spending and taxing, its borrowing and repayment of loans, its issue of new money, and its withdrawaal of money, shall all be undertaken with an eye only to the results of these actions on the economy and not to any established traditional doctrine about what is sound or unsound.” So we look at the EFFECTS and IMPACTS of government spending. When we reach full employment and inflationary pressures are beginning to develop, this is the constraint faced by a government and it should reduce its spending and/or increase taxation to diminish demand. We’re interested in full employment and economic prosperity; we’re not auditioning for the role of finance minister in Zimbabwe, as many of our critics allege.
Thanks Scott, and welcome!
I’m just stumbling around in the dark here, trying to get my head clear on these issues. Like everyone else.
2. (Starting with the easier stuff first). I agree. But I think they already have the accounting about right, from what I’ve seen, and it’s time to move on to the price level. (Except for one pernickety criticism: I wish they were more picky/precise about the tripartite distinction between: quantity demanded, quantity supplied, and quantity actually traded.)
1. I’m not following everything you say here. (My fault, your fault, whatever.) I’m going to pick on just two things you say that I think I do understand:
“Also, whether the govt sells bonds or “prints money,” the deficit has raised the net financial assets of the non-govt sector . . that’s just accounting 101″ But has it raised the net financial liabilities of the government sector (i.e. Treasury + central bank) by the same amount? If bonds pay interest, and money doesn’t, then money isn’t really a liability in the same sense that bonds are. And if the non-govt sector internalises the government’s liabilities, and foresees that it will be required to pay higher taxes to pay the interest on debt, then we get the result that central bank money (assuming it has a monopoly) is net wealth, and bonds aren’t. But then what I have just written depends on whether those future interest payments are indeed financed by tax increases or by printing more money. Are we in a “Ricardian fiscal regime” or a non-Ricardian regime? In other words, in a game of chicken between a government which wants to pay its soldiers without ever raising taxes, and a central bank governor who wanted to control inflation, who would win? In Zimbabwe the answer is Robert Mugabe. In Canada I think it’s the Bank of Canada.
It’s only in cases like that game of chicken that it might matter where the printing presses are located. But even then, Robert Mugabe could just threaten to shoot the governor, regardless of where the printing press is located.
By the way, the governor of Zimbabwe’s central bank, and the old Reichsbank, also said they were only printing currency to meet the demand. And it’s true what they say. But that doesn’t mean they didn’t cause the hyperinflation, or couldn’t have prevented it, unless you say (correctly for at least Zimbabwe) that they couldn’t do anything but print money, given the government would force them to do so to avoid unpaid soldiers or default on the debt.
Marshall: welcome! I was writing my last comment without seeing yours. So the Zimbabwean references weren’t a response to your comment. Yes, you would only be auditioning for the role of Zimbabwean finance minister if you overdid it. (Though what counts as “overdoing it” is not always clear except in hindsight). Will return later. Dunno when.
Nick:
Very glad to see you take on Chartalism (which I am sympathetic to).
You are correct that when the Govt spends, it increases its liabilities just as it increases the non-Govt sectors net financial assets. But think about what “liability” means for a currency issuer, it’s just the financial assets it has paid out to the non-Govt sector. Whether it is currency or debt does not matter, it is still a liability in an accounting sense, although it’s best thought of as paid-out financial assets.
If the non-Govt sector believes it will need to pay higher taxes in the future, or chooses to increase its savings for any reason, then that just means the Govt can run even higher deficits, as that spending will be absorbed into private sector savings, and not “chase goods”, thus causing inflation.
It is true that paying interest on debt can be inflationary, as it forces the Govt to spend. This gets worse when you have CPI linked debt instruments, like TIPS, where the Government is forced to pay out more and more the higher inflation is. If unemployment benefits are “automatic stabalizers” since they increase deficits at times of high unemployment, any CPI linked obligations (TIPS, social security, etc.) are “automatic destabalizers” in that they increase deficit spending during inflation.
PS. At some point, a chartilist may point out the impotence of M in your model because banks do not lend out reserves.
Winterspeak – excellent comment! Even I am against the issuance of TIPS. Someone hit upon the idea that TIPS will reduce government payments since nominal bonds’s yields have a risk-premium built into it and that governments should balance budgets etc and TIPS will help them do so etc etc etc.
Nick,
We all have our views as to what’s important.
Here’s mine on the most important sentences in what Marshall Auerback wrote:
“When we reach full employment and inflationary pressures are beginning to develop, this is the constraint faced by a government and it should reduce its spending and/or increase taxation to diminish demand. We’re interested in full employment and economic prosperity; we’re not auditioning for the role of finance minister in Zimbabwe, as many of our critics allege.”
This responds in boiler plate fashion to your point about the government net position that is the converse to the non government’s net financial asset position. The answer is that it doesn’t matter as a piece of analysis in itself. It only matters that policy makers be vigilant about inflation pressures that may come about from real economy forces.
The point is that Chartalism lifts the illusory belief in financial constraints per se. There are no financial constraints per se on a currency issuer. The only legitimate constraints are those that are self-imposed by a currency issuer on its deficit policy, due to its judgement about the inflation risk arising from real economy capacity constraints. Once that risk and those real constraints become binding, the policy maker responds by adjusting deficit policy. E.g. lower expenditures and/or higher taxes (other things equal) would be a contractionary response to an inflation threat from real economy capacity constraints. The response is driven by real economy considerations; not by reflexive thinking that deficits and debt are “too large”.
JKH,
One minor point. Governments all over have to respond if there are inflationary pressures from the demand side. However, they get into fiscal austerity thinking deficits pass on the debt to the future generation. For example, if there is a fiscal expansion in one year, most probably you may not see it in the next year even if there is no perceived risk to inflation. (Japan is an exception – or maybe was before the new government took over). So they may reduce the primary deficit. So there are two worlds in the Chartalist description – one is the austere world we live in where government don’t see it and media and economists keep criticising and a less austere world in which the deficit is just an accounting number and the primary deficit should be such that it doesn’t lead to price rises.
Hi Nick
Both Winterspeak and JKH have responded to your reply as I would have. I would only add that the distinction from the govt’s perspective is b/n liaiblities that require no debt service and those that do. As I said originally (and as Winterspeak repeated), the latter are actually more “inflationary” (and particularly so when tied to the CPI . . . it appears we already have something we all agree on!).
Best,
Scott
Nick,
My selection from Scott’s original post would be the following sentence:
“The federal govt ALWAYS spends by crediting reserve accounts of banks and recipients’ deposits.”
This is at the very heart of Chartalism’s operational and accounting focus. It positions bank reserve creation at ground zero for the fiat monetary system. It is the operational/accounting reason for taking a consolidated view of a government/central bank balance sheet. And it is the launching point for a liability portfolio management approach to the government/central bank balance sheet. One shouldn’t get hung up here on small think ideas about the definition of a “liability”. The use of the word in this case as a portfolio descriptor is more powerful than a debate about what a liability is and what it isn’t. The portfolio management approach decides how to allocate source reserve creation into reserves, currency and debt. The main allocation is debt. Government debt is an interest rate management tool used in the allocation of pre-existing surplus reserve liabilities – not a source of “financing” in itself. Government spending creates the reserves. There is no financial constraint per se. Finally, this liability management perspective is also the launching point for potential financial system reforms based on a Chartalist understanding of the the existing infrastructure. E.g. it is possible to conceive of a system in which the central bank policy interest rate is set systematically and permanently to zero, no debt is issued, liabilities are left in reserve form, and inflation risk is managed through appropriate deficit adjustment.
And welcome Winterspeak!
Yep, I agree with your point on automatic stabilisers and indexed bonds.
Just remembered one point we all need to keep clear: seigniorage is real (sorry for the pun), but it’s small beer empirically. Suppose non-interest bearing central bank money is (say) 5% of GDP (about right in normal times, I think, for countries like Canada). Suppose it increases at 5% per annum (3% for real GDP growth, and 2% for inflation). So seigniorage is 5%x5%=0.25% of GDP. If you increase the growth rate of base money supply by 1 percentage point, that’s additional seigniorage of 0.05% of GDP, and that’s assuming no inflationary consequences that would raise the opportunity cost of holding non-interest bearing money and so lower the quantity of real money demanded (or a perfectly interest inelastic demand curve for money).
And let’s remember another thing: the demand curve for non-interest bearing money does slope down, and it does have a Laffer curve (and Zimbabwe went over the top of their Laffer curve). So there is a maximum amount of real seigniorage revenue that governments can collect.
It’s not infinite.
What those points mean is that while the ability to print money certainly does relax the long run government budget constraint to some extent, there is still a limit to the government’s ability to finance a real deficit in the long run. (Though it can finance any nominal deficit forever, however large.)
BUT (and this is where monetarists and Chartalists agree I think), what I wrote above only comes into play once we get inflation responding to money printing. So if there’s no danger of inflation, on the margin, go ahead and print! (Though I expect we have to keep a distinction between money growth causing inflation in the sense that it will make the inflation rate rise above a certain level, vs. the sense that inflation would be higher than it otherwise would have been, which might have been deflation).
I want to return to a point I made in the post, but perhaps didn’t make clearly enough. (Because I didn’t understand it clearly enough).
My theory is that there are a bunch of elves running a currency printing press in the basement of the Bank of Canada. And another bunch of elves running a bond printing press in the basement of Treasury. I admit I’ve never seen those elves, but my theory seems to fit the facts of the Bank of Canada’s balance sheet. According to my theory, the bank of Canada’s elves printed some currency, and used it to buy bonds. That’s how the Bank of Canada came to have a lot of bonds in its basement, and recorded on its balance sheet. If both sets of elves, and both printing presses, are really in the basement of Treasury, and the Bank of Canada has no elves, how did the Bank of Canada get the bonds? And how can the Bank of Canada pay Treasury $2billion per year from the interest on those bonds?
By the way, I introduced the elves just to acknowledge that I know my theory cannot be literally true. But it does seem to explain the facts.
Nick: If I understand what you are saying (a bit “if”), you are saying pretty much what I said here I think:
http://ralphanomics.blogspot.com/2009/10/lets-ditch-keynes-and-monetarism.html
Oh those typos: I meant BIG “if”. Ahrrrrrrgh !!
Nick,
The fundamental error in monetarism argument is that it is not stock flow consistent.
Nick,
I don’t know if you’re just having fun here, or you really don’t understand some of these monetary operations and related accounting.
The Bank of Canada does not buy bonds with currency.
The Bank issues currency in response to demand from commercial bank branches, who issue it in response to demand from their customers.
In turn, the Bank debits the reserve accounts of the commercial banks, who debit the deposit accounts of their customers.
In response, the Bank buys bonds from the market in order to create reserves in order to top up the level of system reserves that were depleted by the currency related debit.
Or did you know this?
If so, why are you describing it the way you do, and what’s your point?
JKH,
Well said!
Nick:
Thank you for the welcome.
“Seigniorage” is a meaningless concept in a fiat money system, and therefore chartalism. The idea is that the Government reduces its real debt burden by inflating away the value of its currency. If you embrace the fact that, in a fiat regime, the Government is a currency issuer, you’ll see that this concept makes no sense, as the Government does not have a “debt burden” it needs to inflate away.
Deficit spending (your elves) just create money and give it to the private sector in exchange for stuff. The Government keeps track of how much it has given out, and it is called the “national debt”, but this is the debt the citizens owe to the Government. If the Government was to collect on its debt, it would drain its citizens of all their savings. This would be pointless and calamitous. And contrary to the role of a currency issuer.
When the money is out there, because of how interest rates are set, another set of elves move the money from a checking account that bears zero interest (reserves) to an overnight savings account that bears more interest (Treasury account). This has exactly the same effect as you moving money from your checking account to your savings account in the later afternoon, and then moving it back the next morning (ie. none).
Suppose the Treasury elves went on strike and refused to move money out of reserves. What would happen? It would certainly not constrain the deficit spending elves from deficit spending. But as the banking system would have no drain for its extra reserves (just as if all banks decided to only allow checking accounts, and not savings accounts) it would result in overnight interest rates being zero. Big whoop.
In terms of the Chartalist view of monetary policy vs. fiscal policy (which is the lens I’m interpreting Nick’s elves story through), here is how each is defined:
Fiscal policy is about the changes in the qty of net financial assets held by the non-govt sector. The net financial assets of the non-govt sector are by accounting definition the total qty of government bonds, reserve balances, and currency, less cb loans (or the equivalent) to the non-govt sector.
Monetary policy is about managing the payments system and setting an interest rate (usually the overnight rate). The definition here is in terms of direct operations . . . there can certainly be intermediate and/or policy targets, if one wishes to think that way, but here we’re talking about variables directly and immediately under the cb’s control. CB operations to manage the payments system and achieve the interest rate target do not alter the qty of net financial assets held by the non-govt sector. For instance, open market operations are asset swaps that change the relative qty of reserve balances and Tsy’s in order to achieve the overnight target. CB loans . . either to achieve an interest rate target or promote smooth operation of the payments system likewise add both an asset (the reserve balances) and a liability (the loan or overdraft) to the non-govt sector’s balance sheet.
Thus, the Chartalist view of both is the near opposite of standard macro textbook theory, since fiscal policy is about the quantity of “money” (defined here as the qty of net financial assets for the non-govt sector) and monetary policy is about the interest rate (at least the overnight, risk-free rate, though it could directly set the entire term structure of rates for default-risk free liabilities if so desired, and has in the past).
From these definitions, the monetarist helicopter drop would actually be a fiscal operation, whereas the TARP was a monetary policy operation (to the extent that the financial assets acquired by the govt sector are not worthless).
Best,
Scott
JKH: I am having fun, but mainly I’m trying to make a serious point. I didn’t understand the accounting, but I understand you when you describe it, and I think it’s close to what I would eventually have figured out if I had tried to figure it out.
I know that’s it’s highly improbable that the Bank of Canada would ever buy bonds with currency, just as I know that it’s highly improbable that there are elves in its basement. But at the end of your description, when we look at the balance sheets, I think it is exactly as if the Bank of Canada bought bonds with currency. $100 more bonds in the Bank’s vaults, and $100 more currency in public hands.
You want me to zoom in; I want you to zoom out. And I can only try to come to grips with the differences between our perspectives by using metaphors, like “zoom”, “elves”, “currency” (for all BoC liabilities) and “printing presses”.
Ralph: welcome. I skimmed your post. I think it’s the same policy prescription. But I’m really trying to see if there are substantive theoretical differences.
Ramaman: whoah! Where is it inconsistent?
“But at the end of your description, when we look at the balance sheets, I think it is exactly as if the Bank of Canada bought bonds with currency. $100 more bonds in the Bank’s vaults, and $100 more currency in public hands.”
Yes, true enough. If the non-govt sector desires to hold currency, there is ultimately an asset swap (or a few of them, as there is an intermediate step with reserve balances) changing relative qty of Tsy’s and currency in circulation. But there has been no change to the qty of the net financial assets held by the non-govt sector.
I think Winterspeak was right above . . . this is headed toward a discussion of what “M” is.
I’m hopelessly outnumbered! (And my gf beckons.)
Just one last point: Winterspeak: When I said “seigniorage” I meant dM/P. I think of the “inflation tax” as (dP/P).(M/P). They are only the same if M/P is constant. The inflation tax on nominal bonds is something different again.
Back tomorrow sometime.
Nick,
You look far from convinced but let us make a story.
Let us take some people in an island with good resources, infrastructure and buildings. There is no money because it is a new country. Let us say that the island has enough fruits people can survive on. They just don’t want to survive on fruits and want to have a civilization. They form a government and it is decided to form a currency and that this will be used in paying taxes. Where does the initial money come from ? Let us also throw in a central bank and commercial banks. A standard textbook will say that the government prints $x and then person A deposits and then it multiplies. Please note if this is what you think it is, the next private sector has only $x to work with. Because everything else is both an asset and a liability. The private sector cannot create a net financial asset. Can it ? Soon tax payment time arrives. Lots of trouble, isn’t it ? What if the taxes paid comes out to $(x+3) ?
I view seigniorage as irrelevant or wrong headed for a different reason. It’s an economically fictitious internal transfer payment from the central bank to the treasury. Viewed properly as a consolidated entity, the issue is that a zero rate of interest on currency reduces the weighted average interest cost on the consolidated net liability position (reserves, currency, bonds).
“I think it is exactly as if the Bank of Canada bought bonds with currency. $100 more bonds in the Bank’s vaults, and $100 more currency in public hands.”
Yes. The world is exactly the way it is, as if a lot of things happened in theory.
But they didn’t.
And I sort of noticed the asset liability composition before I explained what actually happened. I was zoomed out, before I zoomed in. It was a compound zoom. It’s my specialty.
And that’s the whole point of the emphasis here – i.e. the operations and accounting for how things actually happen.
JKH . . re: seigniorage
Agree completely. And it’s actually LESS inflationary, as instead of an interest pmt to the non-govt sector, now with the bond replaced by reserve balances or currency, the interest payment goes from the Tsy to the CB and then is returned to the Tsy.
Nick . . . enjoy the day! And thank you.
Nick,
First principles. You said:
“Let’s start with the short run government budget constraint.
G-T +iB = dB + dM
Where G is government spending, T is taxes, i the interest rate on B, the existing stock of government bonds, dB the change in B, and dM the change in the stock of central bank money. Fiscal deficits must be financed by issuing bonds, or money, or some combination of the two.”
Chartalism rejects the idea of a government budget constraint in a fiat monetary system. It considers the above equation to be an ex post accounting identity; not an ex ante constraint. There is no financial constraint, as explained earlier.
And Chartalism rejects the idea that fiscal deficits are “financed” by issuing anything. Deficit spending creates bank reserves as a net government/CB liability. No “financing” is required. The government can exercise an option to issue bonds in order to manage the bank reserve position, and in doing so control the associated trading range for the short term risk free interest rate relative to the CB policy interest rate target. When it does so, the money to pay for the bonds already exists in bank reserve accounts.
This is not semantics. It is cause and effect as reflected in monetary system operations and accounting.
Nick’s Budget constraint is a poor man’s government budget constraint equation. If we separate the central bank and government as different entities, the government budget ‘constraint’ is
G – T + iB = dB
If the private sector (minus banks) buys a fraction x percent of the debt then,
(1 – x)(G – T) = dM where M is deposits. Banks buy the remaining and their assets is up by x(G – T) – no gains as the liabilities is also down by the same amount through the reduction in deposits. Reserves go back to the original value.
Increase in the demand currency is a totally independent event
While I agree that the value of fiat money is ultimately underpinned by the likelihood of government recapitalisation for the central bank if necessary and the willingness of the government to accept money in settlement of tax liabilities, the idea of rearranging the government budget constraint to infer unspoken motives for public sector financial operations seems too far fetched to me. And the suggestion that governments should arrange more of their borrowing through the central bank seems dangerous, even if it does efficiently cut out some intermediate steps. Central banks have been made independent of governments because, in the light of bitter experience of the alternatives, even governments had to concede that independence works best, and I would argue that a major cause of the financial crisis was the subversion of that independence through the (re)appointment of governors like Greenspan and Bernanke.
By the way, partly because of this emphasis on central bank independence, I would say that, in the short term, seigniorage is not the change in M, but the difference between the interest on the central bank’s liabilities (no longer zero) and on its assets, less the running costs of the central bank.
Chartalism above all is an operational and analytical discipline that makes mincemeat out of a loose goose of economic theory about the monetary system.
From an operational perspective, the central bank recapitalization issue is irrelevant, as is the operational separation of the central bank and the government treasury.
Central bank independence is a policy issue but an operational mirage.
Nothing of operational substance prevents the institutional and accounting consolidation of the government treasury and central bank. Conversely, the existing institutional structure obfuscates the operational reality of a fiat system.
The question is the integrity of policy – not false operational boundaries.
Chartalism questions the very guts of the relationship between fiscal and monetary policy to the point where, as Scott points out, it essentially inverts their interpretation. Following operational consolidation, the question becomes what are the rules for portfolio management and policy. Inflation risk becomes a unified policy concern.
Within a unified institutional structure, the question for Chartalists would be who has the authority to change the policy interest rate, should fiscal management of inflation risk somehow fail.
“the idea of rearranging the government budget constraint to infer unspoken motives for public sector financial operations seems too far fetched to me.”
That’s not what’s happening. The GBC is simply being understood as it actually works–as JKH said, it’s an ex post identity, not a constraint. If you can demonstrate operationally or via accounting that it works differently, I’m all ears (or eyes, I guess).
Rebel: ” …I would say that, in the short term, seigniorage is not the change in M, but the difference between the interest on the central bank’s liabilities (no longer zero) and on its assets, less the running costs of the central bank.”
These two definitions are compatible. Ignoring running costs, and interest on M, dM/P equals the present value of the change in the bank’s interest earnings from issuing dM.
More later.
Rebel, JKH and Scott are exactly right. The way I usually like to express it is to simply say that while the government (CB and Treasury consolidated) has real constraints it has no nominal constraint.
Calling it a “budget constraint” just reflects the fact that for private agents who can’t create money the analagous identity does constrain them. The difference is basically just the asymetric redeemability generalized.
The two definitions are not compatible in the “short run” (your choice of words), Nick. Which is how these things tend to be done, hence central bank / government tension in some European countries where the central bank has large gold holdings.
Scott, I mean ideas like that the government issues debt to support monetary policy. This may be true in the sense that, for monetary policy reasons, monetary financing is prohibited by law so that the government has to borrow to spend in excess of taxation, and true in the sense that the government’s treasury or debt management agency generally informs the central bank of its intentions, which the central bank allows for in its open market operations to minimise very short term volatility of its targeted interest rate. But any idea that the politicians have supporting monetary policy on their minds when they consider whether to tax or borrow is ridiculous.
I agree that the state (government plus central bank) has no nominal budget constraint, but does it matter when inflationary repudiation of debt is just another form of default?
The idea that seigniorage has a “present value” is fraught with danger and deception. (It seems to be one of Buiter’s favourite ideas; I assume also it’s a standard treatment in academia.) But seigniorage is merely a circuitous (no pun intended) internal manipulation of the fact that zero interest paid on currency (and at times reserves) reduces the weighted average interest cost of the consolidated government/CB net liability profile. Usually, seigniorage turns out for the most part to be an interest differential between bond interest versus currency/reserve interest (less operating expenses). This is because central banks tend to hold Treasury bonds as their main asset in normal times. But this is an indirect view of what’s really going on, due to the institutional and operational split between the government treasury and the central bank.
Chartalism recognizes that the net liability profile of the consolidated entity produces net saving for the non government sector. That liability profile includes reserves, currency, and bonds. Reserves and currency are obviously the most liquid component of that consolidated liability mix. In addition to reduced liquidity risk, reserves and currency as asset and money classes enjoy an advantage in terms of interest rate risk. Bonds are subject to nominal present value interest rate risk. Assets that pay either zero interest or interest that only resets according to policy rate reset, carry no interest rate risk in terms of such present value.
The favourable liquidity and interest rate risk characteristic of reserves and currency (including risk free immediate encashability for goods and service and/or other financial or real assets) means that any interest rate “risk premium” that should “in theory” be warranted for reserves and currency should be less than the same risk premium for bonds. In fact, one can argue that the reduction in the risk premium is such that it warrants a zero rate or at most a short term risk free wholesale rate (in the case of reserves when reserves are used for unusually discretionary central bank liability expansion, as is the case now).
This in turn means that the assignment of a present value discount rate to the interest rate differential between bonds and reserves/currency is very problematic. The difference is the result of an asset whose liquidity and interest rate risk category is higher than the liability. Separate discount rates apply to each considered separately. That makes even an academic calculation of the present value of seigniorage problematic, because the choice of a single discount rate for such a cash flow is problematic. This is always the case when attempting to do a present value calculation on cash flows that are the net result of assets that are of a different risk class than liabilities. And so it’s very doubtful that it constitutes even a useful academic exercise in the case of “central bank seigniorage”.
(The greatest abuse of such a calculation has to be the horrendous concoction of the idea of “dark matter” on the US net international investment position, somewhat popular a few years back. It was a kind of “foreign sector seigniorage” calculation for the US net international investment income differential. “Dow 36,000” used a similar risk ignorant discounting methodology; that idea didn’t work out so well. The stock market per se is a different beast. Such net risk valuations are essential in order to set trading prices. But I haven’t seen many “prints” in “seigniorage” trades lately.)
Chartalism conversely takes an implicit accrual accounting view (as opposed to present value view) of the interest cost of the consolidated net liability profile, and what the effect of that interest cost is over time in terms of inflation or deflation risk. E.g. high interest cost is inflationary, other things equal, because it puts more income into the non government sector. One of the fundamental advantages of deficit provision of net financial assets to the non government sector in the form of reserves/currency is that the lower interest cost is non-inflationary in terms of non government income, which means tax policy can be relaxed from what it would be otherwise without that low interest component.
Rebel: “But any idea that the politicians have supporting monetary policy on their minds when they consider whether to tax or borrow is ridiculous.”
Agreed. Nobody ever said anything remotely suggestive of this, by the way. The point is that when the fed govt runs a deficit, either the Tsy or the CB must sell bonds or the overnight rate falls below the target to the rate paid on reserve balances (assuming the target rate is above the rate paid . . . obviously now that’s not the case, which is why RBs can be so much higher than banks desire to hold at the target right now). There is no option to “monetize” under those circumstances since bond sales by either are monetary policy operations to support the interest rate target.
“I agree that the state (government plus central bank) has no nominal budget constraint, but does it matter when inflationary repudiation of debt is just another form of default?”
Of course it does. It’s probably been said almost 5 times just in this thread (maybe more . . . didn’t keept track) that inflation is the true constraint. I don’t know how that always gets missed.
Well, I’ve read through the comments. Not sure with some whether I disagree with them or just don’t understand them. Rather than try to address them point by point, I’m going to summarise what I now believe:
The govt/CB has a de facto monopoly on issuing certain types of money (especially currency), and this enables it to earn monopoly profits. We can think of those monopoly profits as “seigniorage”.
A counterfeiter who can print currency undetected steals some of those monopoly profits. If he prints a real quantity dM/P annually, he can use it to buy more goods (increase G), work less (reduce T), or pay off his debt, and reduce his interest payments (reduce B/P and hence iB/P).
We can either consolidate or not consolidate the govt and CB budget constraints. Let’s consider both in turn.
If we think the CB has a different objective from the govt, and doesn’t always do what the govt wants it to do, then it may make sense to separate the budget constraints. In this case, it matters whether it is the govt or the CB that has the power to issue currency (or whatever the money is over which it has de facto monopoly power). If the CB has the power to issue currency, and the govt has the power to issue bonds, I can make sense of the world. The Bank of Canada now has the wherewithal to purchase bonds (which explains how it owns bonds), and has the power to earn monopoly profits which it can remit to the govt annually. If you tell me that it is the govt and not the CB that issues currency, I can make no sense of the world. How did the Bank of Canada get the wherewithal to buy all the bonds it owns? How can the Bank of Canada with a tiny net worth earn the $2B profits it remits to the govt annually?
Alternatively we can consolidate the CB and govt budget constraints. In which case it is irrelevant which one issues the currency, because they are both the same entity.
Working with a consolidated balance sheet from now on:
As Adam P says, the govt has no binding nominal constraint.
As Adam P says, the govt does have a binding real (i.e. price-level-adjusted) constraint, BUT:
i) seigniorage gives it an additional source of revenue that is not available to normal people.
ii) AND (this is where I line up at least in parallel with the Neo-Chartalists, if not side-by-side), if inflation is not a danger, and until you hit the point where inflation is a danger, seigniorage is UNLIMITED. Or, to say the same thing another way, inflation eventually getting “too high” if you push it “too far” is the ONLY thing that limits seigniorage. Just to be precise, “too high” can be interpreted politically (as in Canada’s decision to target 2% inflation), or ultimately it can be interpreted as the top of the seigniorage Laffer Curve, the point where the demand for currency is unit-elastic. And “too far” means “past the natural rate” (which translates roughly as “full employment” in the monetarist-Neo-Chartalist dictionary).
And if any critic of the Monetarist/Neo-Chartalist position on this point says “But won’t printing money cause inflation?” our proper response should be: “It certainly will cause a higher rate of inflation than would happen otherwise (at least we hope it will), and we want that to happen, because what would happen otherwise would be deflation (or inflation less than 2%)!”
If inflation is not a danger, and until inflation is a danger, a government that finances its deficits by issuing bonds is doing something stupid, sort of like imposing a financing limit on itself where none need exist. (I think Neo-Chartalists are saying the same thing.)
But why do Neo-Chartalists stop there? You aren’t radical enough, you timorous bunch! (Or maybe I’ve misunderstood you.) If inflation is not a danger, don’t just print enough money to pay the current deficit; print a whole lot more to buy back some bonds, so you won’t have to pay the interest on those bonds in future, which means you will be able to increase G and/or reduce T in future years. And keep on doing this till you get to the point where doing any more would make inflation a danger.
JKH: I think I see where you are coming from on the present value of seigniorage point. But look at it this way: if we vary the composition of the total govt debt (defined broadly to include currency etc.), to have more currency and less bonds, then the total interest payments will be lower. And we can define the present value of seigniorage, if we take the flow of interest rate differential times the stock of currency, discounted at the rate of interest on bonds.
Nick,
My comment on seigniorage was a sort of personalized, experimental digression. It has nothing to do explicitly with anything I’ve picked up on Chartalism. I’ve not noticed any emphasis on the subject of seigniorage per se in what I’ve seen on Chartalism. Your explanation of the calculation is the more normal straightforward one of course.
However, your struggle with the issue of how the central bank “earns” seigniorage makes me think my comment was worthwhile, as several aspects of the comment are pertinent to this.
Perhaps Scott Fullwiler can put the subject of seigniorage into Chartalism context more formally and authoritatively.
I don’t have much time to look over the discussion of seigniorage carefully, but my impression is that JKH’s points are in line with Chartalist views. And, as I think JKH and I already noted, seigniorage “income” in a modern monetary system (like in the US and Canada) is essentially about reduced spending on interest to the non-govt sector.
Randy Wray did a paper in 2003 (published in a book edited by Rochon/Rossi titled “Modern Theories of Money”) titled “Seigniorage or Sovereignty” that may be of interest. Here’s the link to the wp version: http://www.cfeps.org/ss2008/ss08r/Tcherneva/Seigniorage%20or%20Sovereignty.pdf
Forgot to note that I’ll try to look the discussion on seigniorage over more carefully later. Thanks again, Nick, for your willingness to engage so openly.
“seigniorage….an additional source of revenue that is not available to normal people”
Interestingly, not quite true. Various pre-paid cards, like the London Oyster card, can capture some of the seigniorage that the central bank (or the government perhaps; in some countries, coinage is a liability of the government) normally receives on the currency held for certain everyday transactions.
What I was wondering, Scott, is whether inflation should be the constraint. Is it possible that the risk premium that the government pays for the possibility of inflationary repudiation is no less than for the possibility of a simple non-inflationary default? In other words, if the state intended to run up as much debt as it could and then not pay, would inflationary repudiation be its optimal strategy?
Nick,
Thinking about your particular questions:
The central bank responds to non government demand for currency by offering a supply of that currency on demand. (I think you economists might refer to that as an infinitely elastic supply curve at a zero rate of interest, but I probably embarrass myself here.) This is one reason I ran through the accounting for this earlier.
The central bank controls the level of reserves in the banking system (including the current discretionary case of excess reserves created via “credit easing”.)
The level of “seigniorage” as defined in the context of a deconsolidated central bank balance sheet therefore is a function of the bank’s non discretionary provision of currency and discretionary provision of the level of reserves.
The central bank controls the market trading response (within reasonably contained volatility bounds) to the short term policy rate.
The central bank can change the short term policy rate at any time.
Therefore, the central bank may have some indirect influence on the demand for currency in the long run, but I would be careful about this.
The question of who issues the currency is not relevant to the consolidated operations and accounting for the fiat system. It happens to be the central bank. Again, with the same accounting example for the issuance of currency as noted above, I thought we already answered the question of “where the money comes” from to buy bonds.
The question of who issues the currency is relevant in a policy sense within an institutional framework of competing political forces, where they exist.
On your second last point, Mosler and others recommend no bond issuance as you imagine. Inflation risk is managed through deficit control.
Nick,
If you think the Chartalists are a timorous lot, I suggest you have a close look at the following. You may end up tossing some cookies:
http://www.moslereconomics.com/2009/09/16/proposals-for-the-banking-system-treasury-fed-and-fdic-draft/
(Click on the second big highlight)