Who will buy the bonds?

"The government needs to sell bonds to finance an expansionary fiscal policy. But who will buy the bonds? What happens if nobody wants to?"

What we ought be be concerned about is the exact opposite. The fiscal policy will be more likely to succeed if people do not want to buy the bonds.

Suppose, just as an example, that the proposed fiscal policy is a bond-financed tax cut, holding government expenditure constant. And suppose, again just as an example, that Ricardian Equivalence holds exactly. In this example, there's a simple answer to the question: "who will buy the bonds?". Each person receiving a $100 tax cut (or transfer) will want to save the whole of that $100, and buy an extra $100 of bonds. The extra supply of bonds creates its own demand. And yet this example is precisely the case where fiscal policy will fail to stimulate aggregate demand.

An increased demand for bonds means extra savings, and we don't want that to happen. We don't want an increased supply of government bonds to be met with an increased demand for government bonds. We want to create an excess supply of government bonds at existing levels of interest rates, prices, and incomes. We want to create a disequilibrium in the bond market, that will force interest rates, prices, and incomes to change. We want to create a disequilibrium in the bond market that will spillover into a disequilibrium in the market for newly produced goods.

Let's take what is conceptually the simplest example: the government prints bonds and gives them to people. (This is equivalent to helicopter bonds or a bond-financed tax cut). If people just hold the extra bonds, that is the end of the story. Nothing else happens. We hope that people will want to get rid of the bonds. We hope that the supply of bonds does not create its own demand.

We want people to try to get rid of the bonds, by trying to sell them, and trying to buy goods.

Bonds are not a medium of exchange, but money is. People cannot directly buy goods with bonds. If an excess supply of bonds is to translate into an excess demand for goods, it must first create an excess supply of money. There are two ways this could happen.

First, the excess supply of bonds could push down bond prices, which means push up interest rates on bonds. And the higher interest rates could cause a fall in the demand to hold money, creating an excess supply of money, and an excess demand for goods.

Secondly, the Central bank might buy the excess supply of bonds, in order to prevent the rise in interest rates. When it buys bonds it sells money. That creates the excess supply of money, and an excess demand for goods.

Now let's consider a bond-financed increase in government spending on goods. As a first step, the government needs to sell $100 bonds for $100 money; in the second step it uses the $100 money to buy $100 worth of goods. The net result is that the private sector holds the same amount of money, but $100 more bonds. If the private sector was willing, at unchanged interest rates, prices, and incomes, to hold the same amount of money and $100 more bonds, that would mean the fiscal policy would fail to stimulate demand. Because the only way the private sector could hold an extra $100 of bonds is if it saved an extra $100 of its income by reducing its consumption by $100. But if a $100 increase in government demand for goods lead to a $100 decline in consumption demand, the net effect on aggregate demand would be precisely $0.

Again, what we should be scared of is not that people won't want to buy the bonds, but that they will want to buy the bonds (at existing interest rates, prices and incomes). We want to create a disequilibrium on the bond market. We want fiscal policy to force interest rates, prices, and ultimately incomes to change.

In general, we should be more worried that people will want to buy and hold an extra supply of bonds than that they won't want to. Ideally, for fiscal policy to be most powerful, the private sector would refuse point blank to hold any more government bonds, at any rate of interest, price level, or level of income. In this case the central bank would be forced to buy all the bonds, so government expenditure would be money-financed. And if people refused point blank to hold any increased stock of money as well, we would be incredibly lucky. Because the extra $100 stock of money would create a $100 excess supply of money, which people would try to spend out of existence, again and again, and the hot potato would pass from one hand to another indefinitely, making the fiscal multiplier infinite.

But I think there is one important exception: if the fiscal authority in question does not have its own money and central bank, things are different. If we are talking about fiscal policy in a Canadian province, or a Eurozone country, we might be worried the other way. I will deal with that question in a later post.

Comments

[Don the libertarian Democrat]

Nick, Have you seen this blog post many people are talking about:

http://blogsandwikis.bentley.edu/themoneyillusion/?p=349

What is it about Buiter's position that makes it so unacceptable?

http://www.nber.org/~wbuiter/helijpe.pdf

It seems to me that your position is similar. What am I missing?

[Nick Rowe]

Hi Don: I have read Scott Sumner's post (and read through the other posts on his blog as well). I like it. My view of the world is similar to his. A couple of months back I wrote a post advocating price-level path targeting. And I have pushed for a more aggressive unorthodox monetary policy. Same sort of thing as Sumner. (Not sure I think that eliminating the 0.25% interest on reserves will make a big difference though, but sure, why not?)

I'm not sure what you mean about Buiter's position being unacceptable? Do you mean: "why doesn't the Fed do a helicopter increase in the money supply?"?

The key to Sumner's view, and Buiter's paper, is that it is by influencing expected future monetary policy that we can increase aggregate demand, and escape the recession. And the problem is: *how* to influence expected future monetary policy?

A public commitment to a price level path (or nominal GDP path) target would help. But will it be credible?

My own post "Bernanke should bet on inflation" (or whatever the title was) was on the same theme.

Is that what you were asking?

[Nick Rowe]

And yes, his talking about creating an excess supply of money (a disequilibrium) is in line with my way of thinking. It is one of the things I picked up from David Laidler, who was my supervisor. Good monetarist thinking.

[pointbite]

We've had excess demand for many years and now we must restore balance to the economy, sometimes doing nothing is the best option. Why must we try to re-inflate demand if that demand was excessive? (proof: debt levels relative to income). I found it difficult to get through this post, it's basically Russian roulette, if you get the formula just right perhaps the party goes on a few more years... if not, you risk destroying confidence in your currency. Is it worth the gamble?

[Nick Rowe]

pointbite: A rising debt (or rising debt/income ratio) does not mean demand was excessive. Rising debt means that some people (debtors) have been spending more than their incomes, and other people (creditors) have been spending less than their incomes. For every $ borrowed there is a $ lent.

[himaginary]

My question is, then, why not let central bank buy bonds directly from the government? As I commented in earlier post, I don't see any merit in making disequilibrium in bonds market and let interest rate climb even if temporarily. Rather, such disequilibrium seems to do a lot of harm.
FYI: In Japan, BOJ bought bonds directly from the government during the WWII. That action deemed to have lead to hyperinflation after the war, and that experience is one of the main reasons why BOJ is so sensitive about inflation even now when the problem is deflation. I don't think central bankers in US or Canada have such psychological binding.

[ClaudeB]

I understand Canadian provinces and member countries of the Eurozone have no control over monetary policy, but there are some (limited) alternatives; some jurisdictions still have tools to absorb bond issues without resorting to private savings. For instance -- and regardless of its current problems -- the Caisse de dépôt et placement du Québec could stock up on Quebec bonds up to $4G/year, with the deposits it receives, such as RRQ contributions.

The Government would get its $100 of money to spend on goods with $0 decline in consumption, since the bond market would not buy the bonds. Et voila!

[pointbite]

Nick: Those creditors are in Asia, America as a nation spent much more than it earned. Demand was excessive and the problem of excessive demand will not be solved with more demand. I doubt the government of Canada can do anything to restore confidence in a vacuum. If we can't force Asians to create their own demand, I fear trying to re-inflate an American bubble will only destroy the US dollar and by association many other currencies (we've never before had a global reserve currency backed by nothing, who knows what would happen if confidence was lost). I just find this whole discussion quite distasteful. I would rather we just experience a few years of recession, we've gone soft.

[Don the libertarian Democrat]

It seems to me that the same ideas keep coming up in slightly different forms. I liked Buiter's idea because, if I remember right, it didn't involve issuing bonds. I liked your idea, but for the fact that it's linked to toxic assets. I just felt that we should leave them out of it. I don't like the Fed's QE, because it seems to be hedged in anticipation of future problems.

I read Bernanke's talk on ZIRP, but he seems not believe his own writings. I've also read that he's a big Fisher fan, which led to Bears, but the Lehman decision seems odd then.

So, yes. How many times am I going to read a QE idea trumpeted on blogs, and yet the idea goes nowhere? What is the Fed worrying about?

[Nick Rowe]

Pointbite: one further response to your first comment (I forgot to say this at the time): inflation (above target) is evidence of excessive demand.

Now on to your second comment: I was thinking about Canada, not the US. Canada has had very small foreign borrowing in recent years. Now let's talk about the US. We need to distinguish between demand for American-produced goods, and demand for goods by Americans. When economists talk about "aggregate demand in the US" we mean the former. The economists' term for the latter is "domestic absorption".

So, aggregate demand in the US has been roughly about right, but is now too low. Domestic absorption in the US has been too high in the past. It has caused foreign debt to get too high. But not all debt is foreign debt. Canada for example has had increasing levels of debt, but no (negligible) foreign debt.

himaginary: If the central bank decides to keep interest rates fixed, and the private sector does not want to hold the extra debt, then the central bank will buy all the extra debt, and interest rates won't really rise, even temporarily. I was just working through the example one step at a time, so there seem to be delays between steps, which makes it look like the rate of interest rises temporarily. Just a professor's trick, to tell the story slowly. It doesn't really happen like that. It all happens at once. Sometimes we use the code words "*incipient* rise in the rate of interest" to mean "it doesn't really rise, but would rise if the central bank didn't see it coming and prevent it from rising". But yes, it might as well buy the bonds directly from the government. And yes, one of the problems with getting the BoJ to do an aggressive enough monetary policy, as I understand it, was a constitutional provision that disallowed such purchases, reflecting postwar fear of inflation. The BoC has no such constraints. Nor does the Fed.

ClaudeB: I'm going to wait for the provincial/Eurozone post!

[Nick Rowe]

Don: that's a good question. I don't really have the answer. Here are some part answers:

It has been tried to some extent. The Fed's balance sheet is much bigger. The BoC bought mortgages, and lent against weird colateral. See my last post on "what next for the BoC".

It is difficult to know how to make a credible commitment for future monetary policy. How do you set the printing presses running, and disable the "reverse switch"? Especially when people know you might want/need to use the reverse switch if inflation gets too high.

Nobody has a clue about how big the policy would need to be. How would we know when we've overdone it? Sumner's posts have been helpful here: his answer is to look at private forecasts, or market forecasts, for nominal GDP (or CPI). But then there was that paper (by Bernanke? Blanchard? somebody beginning with B) saying you can't target a forecast without disappearing up your own orifice. (Sort of self-referential paradox).

Fear of central bank balance sheet losses. (I think they are overblown, and might actually be good for credibility, as in my "betting" post.

Fear of "instrument instability". We keep slowly moving the policy lever, more, more, more, and then suddenly it's effective, too effective, help! We're heading into hyperinflation! reverse the levers fast! Damn, too much!

Fear of the unknown. Orthodox monetary policy is an M16, but it's run out of ammo. Fiscal policy is Grandfather's musket. Unorthodox monetary policy is a nuke. Nobody knows the critical mass, or how big the bang will be.

[jimbo]

Interesting. I've tried to explain this here before, but I don't seem to get through.

You are working in the wrong paradigm. Your question is nonapplicable, silly even, in a floating exchange rate fiat currency world like the one we live in.

Who will buy the bonds? Anyone who wants a non-zero interest bearing instrument rather than a non-zero interest bearing instrument (namely reserves). There are no other choices. Govt. spending adds reserves to the system. The only way those reserves can be drained from the system is via taxation or by issuing bonds. There is no other place for the money to go. An individual can take his money and choose to buy, say, a corporate bond or stock or something else rather than a govt. bond, but there is no way the system as a whole can rid itself of excess reserve balances without either the fiscal or the monetary authority acting to drain them. The act of selling bonds s not, in fact, a "borrowing" operation at all, since operationally the money must be spent into circulation BEFORE it is drained by the bond sales.

[jimbo]

Oops. Make that 2nd "non zero interest bearing" into "zero interest bearing"...

[Don the libertarian Democrat]

Nick,

For what it's worth, I'm a big fan of Econbrowser, but you do the best job of explaining and commenting on current economic issues, and throwing out interesting ideas.

Don

[Nick Rowe]

Don: Thanks very much for saying that. I really appreciate hearing it! Nick.

Jimbo: For some reason (maybe my fault) it's still not getting through to me. Let's keep trying.

At the back of my mind, I had a fiat currency, floating exchange rate model in the back of my mind when I wrote that post. Let's just make it a closed economy, to keep it simpler still, and find out exactly where we either disagree, or perhaps misunderstand each other.

I agree with everything you say in the third paragraph. OK, I might quibble over what happens first. Does the government first borrow money from the central bank, spend it, and then the central sells bonds to get the money supply back to where it was? Or does the government first borrow the money from the public, and then spend it? But I don't think it matters to my argument. Or was this where you believe we disagree?

But the bit about the system as a whole being unable to get rid of money was exactly what was in my mind when I talked about the "hot potato" of an excess supply of money.

And I assume that we start in equilibrium in asset markets, with people already holding the desired mix of bonds and money.

Where do you think we disagree?

[JKH]

Nick,

I don’t want to misattribute or falsely associate the comment of jimbo above, but it is very reflective of the economic philosophy of one Warren Mosler, documented in great detail at his site Mosler Economics, here:

http://www.moslereconomics.com/mandatory-readings/soft-currency-economics/

If you are on sabbatical, I think it’s well worth spending a little time looking at this, although it is tough slogging.

My impression of it is that it’s a very insightful and detailed look at the operational guts of the banking system with particular emphasis on how government budget transactions can affect banking system reserves.

But in my opinion, Mosler goes too far in extrapolating this knowledge to a larger monetary theory.

As just one example, he says that government budgetary expenditures essentially create the banking system reserves that are required to support tax payments or bond purchases. In that sense, expenditures fund revenues. The essence of the idea is that government expenditures have the potential to create an overdraft position in the government’s account at the central bank. An overdraft is essentially an extension of credit. If that overdraft is left uncovered, the government has funded its expenditures by printing money, just as does the central bank in its normal monetization operations. The government in this sense can be viewed as an extension of its own central bank, and taxation and debt issuance can be viewed as extensions of monetary sterilization.

This is true as an explanation of how the system can work. But Mosler seems to extend this to the notion that the system does work this way, in the sense that expenditures tautologically precede taxes or debt financing.

That’s wrong, in my opinion. The system can work this way. But there’s also nothing to preclude the operational fact that governments can collect taxes or issue bonds (thereby draining bank reserves) before making corresponding expenditures (thereby replenishing bank reserves).

The result is a grand Mosler theory which he calls being “in paradigm”. Those who don’t understand it are branded as “out of paradigm”. There are other deeper and impressive ideas associated with this theory which admittedly I haven't had the time to explore yet.

Nevertheless, a great contribution is his deep understanding of the operational details of the banking system that is very instructive in linking it with economic theory more generally.

[Nick Rowe]

Thanks JKH. Before reading the Mosler piece, what you say makes sense to me.

Slightly off-topic: do Canadian provincial governments have bank accounts at the Bank of Canada? (I don't think they do). Do they have bank accounts at the commercial banks? (That is what I would assume). What about Eurozone countries?

Even more off-topic (but while I'm asking basic questions): what percentage of the whole Canadian financial system do banks represent? In other words, how much borrowing and lending goes through financial intermediaries (continental model), and how much through stock and bond markets (anglo-saxon model)? I have heard 25%. Is that about right (roughly)? How does Canada compare to other countries?

[JKH]

I don’t believe the provincial governments have accounts at the Bank of Canada. They do have accounts at the commercial banks. I’m not familiar with Eurozone operations.

BTW the federal government holds balances at both the Bank of Canada and the commercial banks. This is a critical opeational feature allowing the Bank the ability to fine tune the level of supplied reserve balances.

I’d really be taking a wild guess on the second question, but a very pertinent subset of it is how much of the Canadian residential mortgage market do the banks hold on balance sheet. There I’d guess in the area of 50 per cent. But I haven't looked closely at such numbers in a long time.

[Mark Dowling]

Nick - what about the scenario where if people don't buy bonds they still park their money? Assuming it doesn't go under the mattress, which I'm betting is happening anyway, isn't it preferable that if the savings rate is going up regardless, people can buy government bonds knowing the government won't use it to pay down debt rather than giving it to the banks who won't lend it out because they're trying to keep their balance sheets looking good?

[Nick Rowe]

JKH: I had a quick read of the Mosler piece. His is a fairly standard Post-Keynesian approach, as far as I can see.

The strange thing is, that even though I am not a Post-Keynesian, I have very few disagreements with his perspective. The disagreement is more methodological. While Mosler seems to be saying "this is the way the world is", I would say "that is one way of looking at the world, but there are others, equally valid, and if we take into account all the feedbacks we end up in the same place anyway". It's more a difference in what order you tell the story. We can start with a liquidity preference endogenous money approach, then go on to consider feedbacks of monetary policy and loanable funds. Or we can start with loanable funds, then go on to consider feedbacks through monetary policy and liquidity preference.

I'm old enough, have been taught by many different profs, with many different perspectives, etc., that I can usually tell the same story from either direction, and end up in the same place, if I don't get muddled along the way.

What he sees as features of the world I see as features of models of the world.

I think you and I basically agree on this, JKH.

Yep. But I thought the BoC stopped using the "drawdowns and redeposits" instruments a few years back? I was so thankful when they did, because I always got it muddled when teaching it.

Mark: Suppose an increased supply of government bonds is met with an increased demand for money (for some reason). I don't think it would make much difference in current circumstances, since the BoC would just swap money for bonds, to give people what they want. But if the supply of money were assumed fixed (as commonly assumed in many textbook models), then bond-financed government expenditure would be highly contractionary, because it would create an excess demand for money.

[pointbite]

Nick, can you realistically stimulate aggregate demand without inflaming domestic absorption? America leaks like a sieve. In Canada we make too many primary goods, if the public had more money they would be unlikely to purchase more nickel and lumber, they would probably buy a finished product from Japan or Germany or China... Sure some of that money may filter back to Canada, but not all of it. I'm still uneasy about the whole micro-managing demand business, it's a little too central-planning for my taste.

[JKH]

Nick,

Interesting about drawdown and redeposit. I wondered about that. You're probably right. Maybe I'm way out of date. It used to be very important.

[Nick Rowe]

pointbite: Hmm. TypePad seems to have lost my original reply.

In normal times it would be easy to do what you want. Tight fiscal policy and loose monetary policy would do it. But now, when it's hard to get looser monetary policy, it's not so easy. But US savings rate has recovered, so maybe domestic absorption is not so much a problem any more?

[jimbo]

Yes, I will admit to being an admirer of Mr. Mosler, as well as otehr Post-Keynesians like Randall Wray and Wynne Godley. Maybe I get a bit shrill about it times, but given the state of the economy and the economics profession these days, I think I little shrillness is called for...

JKH, you lost me here:

"But Mosler seems to extend this to the notion that the system does work this way, in the sense that expenditures tautologically precede taxes or debt financing.

That’s wrong, in my opinion. The system can work this way. But there’s also nothing to preclude the operational fact that governments can collect taxes or issue bonds (thereby draining bank reserves) before making corresponding expenditures (thereby replenishing bank reserves)."

How can the monopoly issuer of a token not issue it before it collects it? Can an airline collect tickets at the gate without selling them beforehand? Can a subway collect tokens at the turnstiles without issueing them? Sure, the fact that ther are savings sloshing around in the system means there's not necessarily a one-for-one relationship between issuance and redemption, but the logic is inescapable. I'm curious how you think it is possible.

I don't know much about the Canadian banking system, but "drawdown and deposit" sounds like the Treasury Tax and Loan accounts that the U.S. gov. uses as a buffer to manage reserves. Is that about right? If so, I would guess that Canada would no longer need them since it went to its 0 reserve policy, since it doesn;t have the same machinations to go through to maintain reserve balances.

[JKH]

jimbo,

First of all, I like Mosler’s stuff a lot. I’m taking my time getting through it all, and therefore I can’t claim to understand or appreciate it fully.

On your point, I would say the following:

The private sector pays taxes via debiting of their bank accounts and crediting of government account(s). The government ultimately gathers those funds at a central point, that being its account at the central bank. This crediting of the government account is mirrored in the debiting of bank reserve accounts by the same total amount. This is the process whereby taxation at the margin drains reserves from the system.

All that is required for this to happen independently of government expenditure is the existence of a sufficient outstanding stock of M1 type balances.

The government is a monopoly issuer of currency and reserve balances. It is not a monopoly issuer of M1 balances. The commercial banking system issues M1 balances, and as Mosler would agree, is not constrained by central bank reserve requirements in doing so. So the M1 balances can be made available for people to pay their taxes before the expenditures that offset those taxes need occur. The system can work this way.

On a cumulative basis, it has worked this way over certain non-trivial periods, in the examples of Clinton and Canada that I have previously noted (i.e. budget surpluses).

It would in theory be possible for a government to collect taxes, and still be in budget balance from the beginning of its existence, with zero outstanding debt. Moreover, central bank liabilities could be invested in private sector assets. The currency that forces its own acceptance through the power of taxation would still be outstanding as it is now.

All I’ve said is that taxation can precede expenditure in arriving at surpluses, balances, or deficits. I’ve not said that currency is not outstanding or hasn’t been issued.

As I suggested to Nick, I think you’re right about “drawdown and redeposit”, and it is comparable to Treasury tax and loan accounts. I’ve haven’t investigated it recently but the precise operation and effect depends on payment systems technology, same day settlement characteristics, etc. etc. That said, the distribution of treasury balances between accounts at the central bank and at the commercial banks is at the discretion of Treasury/the central bank and this goes for the US and Canada. This distribution has been pretty important as temporary funding for the Fed balance sheet.

[jimbo]

JKH -

We might be splitting hairs a little, here. I agree that taxes and expenditures do not have to be precisely correlated. The system you are describing seems to me to be similar to the system in the U.S. after Jackson killed the Bank of the U.S., with gov. liabilities kept in several private banks.

But I guess that's what Mosler gets on about: he's not talking about the way it must be, or can be in every possible theoretical configuration. He's describing how a current, central bank based floating exchange rate currency system works, and the imperatives and policy possibilities that flow therefrom. (He's always struck be as a practical-minded sort more interested in operational realities than theoretical considerations - which can be both a weakness and strength.)

[JKH]

jimbo,

I wrote this just before seeing your last, with which I agree. Mosler is strong on operational reality, which I think is very important. E.g. I was struck right away on discovering his site a few months ago by his bit on the multiplier, which is totally correct, but not understood by many. I just wish he’d tweak his idea of expenditure/taxation “ordering” a bit but I’m not sure how exactly.

Anyway, rather than waste this, here it is, although a bit redundant now:

“How can the monopoly issuer of a token not issue it before it collects it? Can an airline collect tickets at the gate without selling them beforehand? Can a subway collect tokens at the turnstiles without issuing them?”

I don’t see this as necessarily proving Mosler’s base case. Mosler’s argument seems to be that expenditure precedes taxation at an operational level. Governments spend, which increases bank reserves. Reserves are then drained by debiting the banks whose customers are paying the taxes. It can happen this way, but it doesn’t have to.

Assume the airline is analogous to the government. Suppose expenditure precedes “taxation” in this case. The airline spends to pay for the cost of a flight (fuel, labour costs, etc.). It pays by issuing its “monopoly currency” - in this case airline tickets. The system then has excess “reserves” in the form of airline tickets. Supplier/customers of the airline pay their “taxes” to the airline by presenting their tickets as payment for the flight. Reserves are drained.

It can happen this way. But it doesn’t have to happen this way. The order in which the airline spends and the customers redeem is operationally irrelevant. The airline may pay for its fuel bill or its labour costs before or after the flight. It may run a temporary deficit before the redemption of the tickets or a temporary surplus after the redemption. It is not constrained in any way by the notion that expenditure must precede taxation.

The fact that the airline has issued its currency before the currency is redeemed is also irrelevant. This only means that the airline is running a gross surplus with respect to the lead time between the issuance of its currency and the time of the flight or redemption of that currency. In fact, it argues against the Mosler paradigm as a marginal effect, because the result taken on its own is a taxation surplus. But what must be compared against this gross effect is the other side of the transaction, which is the airline expenditure. That can happen with any timing – before the issuance of the currency (strong Mosler), between the issuance of the currency and the redemption of the currency (weaker Mosler), or after the redemption of the currency (non-Mosler). The net effect will determine whether the time distribution is one of deficit followed by a netting surplus, surplus followed by a netting deficit, or “simultaneously” timed balance.