« Advice for new Teaching Assistants | Main | The inefficiency of monopoly: Remembrance Day edition »

Comments

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

Different people have different conceptions of how government debt works. To some, government debt is a junior liability. Money is a senior liability. An excessive public debt (assuming most of the debt is bonds) then results in default risk but no inflation. No AD boost. This is I guess the more traditional view and what most people mean by "bond vigilantes"?

And then there's the fiscal theory of the price level / MMT view that regards money and bonds as equal government obligations, and so an unsustainable public debt causes a rise in the price level rather than an appearance of default risk.

"The bond vigilantes will eventually attack and rescue the US and Japanese economies from recession. But every year they wait before attacking, the bigger that attack will be. And they won't attack until they have grown bigger than we would want them to be."

Oh, yeah?

I observed the attack of the bond vigilantes in the late 1980s. The trigger for the recurrent attacks was unemployment falling below 7% or threatening to do so. I would not count on them to rescue any economy.

Nick: because it would increase Aggregate Demand too much. That would force the Fed to increase interest rates a lot, and that would force the US government to raise taxes and/or cut spending to cover the increased costs of servicing the debt.

Increased demand by definition increases taxes and reduces. That is in absolute measures, not in %. Why do you still say that it would *force* the US government to do anything? I can play golf all day long and budget deficit might even turn in surplus if demand really increases Too Much? What is the problem?

The "max-min" stuff above is really funny :)

sorry guys. but could not help :)

Nick, pls delete if you mind

How could the federal reserve be "forced" to raise interest rates?

One would have to ask: Is the government debt serviceable in U.S Dollars or in the currency of the country who owns it/where the person resides? For example, china keeps their currency low because they get more of their currency for each one of our dollars. Would the difference with the worth of our exports(the rise in income from being able to keep or even lower the price for our exports and still maintain or increase our profit margin) off set much of the problems you refer to?

Nick, please tell me if I correctly understand the issues here.

1) Krugman clearly says he's disregarding inflation. You're arguing that inflation matters if the speculative attack is big enough. Okay...I assume you're talking about the impact of the USD depreciation on the US inflation rate. But are you talking about the direct impact of that depreciation on the US price level? or the indirect impact of a real depreciation on US competitiveness and subsequent inflation via excessive aggregate demand? or both?

2) To get higher interest rates, are you assuming that the Fed responds by ignoring the dual mandate and just targeting inflation (or the price level)? or are you worried that a depreciating dollar itself makes bond traders demand an interest rate premium on the USD as compensation for future expected inflation?

3) Sergei notes (above) that higher AD improves government balances while you note that higher interest costs cause them to worsen. The net effect will depend on the relative size of these channels. What arguments suggest to you that one is larger than the other? I'd note that interest rate hikes start to affect debt service costs only as the debt is rolled over. I think most countries that are hit by the kind of debt service/interest rate spiral you describe (under floating exchange rates) have faced a long period of increasingly steep yield curves and respond by aggressively shortening the average maturity of their govt. debt, making their deficits more sensitive to interest rate changes. Isn't that quite different from what we've been seeing in the behaviour of US yield curves and average debt maturity these past few years?

You know that you're picking a fight with a Nobel Laureate whose prize citation mentions his seminal work on exchange rate crises. If you want to press the attack here, it will probably be in your self-interest to be very clear about your reasoning.

One other thing you might want to think about.....

If the bond vigilantes attack the USD, the US will be the least of your worries. A suddenly weaker US dollar against other world currencies implies a suddenly stronger Euro and Yen (all that extra AD that the US gets is just being transferred from other economies) thereby making the weakest major economies weaker still. The impact on a shaky banking system without a clear lender of last resort is something that would worry me more than the impact on the US.

At the same time, the depreciation of the USD would cause hundreds of billions of dollars of losses on the multi-trillion dollar hoard of USD claims held by the Chinese, I would also doubt that their banking system could withstand that without a massive intervention by the state. A potential banking crisis in Europe is bad enough without another potential one at the same time in China.

Of course, worries about those potential banking crises could make investors prefer....USD!

Simon:

1. both. Too much AD causes too much inflation. That is true in both open and closed economies, though the mechanisms are slightly different in the two cases.

2. both. If AD is too high for any determinate path for the price level, at current real and nominal interest rates, the Fed will have to raise real and nominal interest rates to prevent an eventual explosion of the price level path. Standard Howitt/Taylor Principle. PK already (partly) has that effect in his model, when he includes the Fed's reaction function with r responding to Y. (I would just add inflation to the Fed's reaction function, which he has suppressed for simplicity.)

3. Empirically, the US budget was in deficit even before 2008, IIRC. A plausible assumption is that recovery would mean returning to that same primary deficit, plus the total deficit would be bigger because the debt/GDP ratio has increased since then. Right now the interest rate is below the growth rate of GDP, so the bond vigilantes have no worries whatsoever about debt/GDP as long as r stays below g. But if r rises above g the LR government budget constraint bites, and the government must be expected to run a present value of primary surpluses equal to the current debt. The bigger the debt/GDP ratio, the greater the eventual natural rate of r when the economy recovers (standard OLG model result). And the longer it takes before it recovers, given current deficits, the bigger the debt/GDP grows, and the bigger the jump in r when it does recover.

Having long average maturity of the debt reduces rollover risk and gives the government more time to respond to a rise in r by raising taxes and cutting spending. So the risk of a sudden attack is lowered. But the government must still eventually respond.

Bottom line: we don't know what the maximum sustainable debt/GDP ratio is. It depends on the government's ability (both economically and politically) to run primary surpluses, and on the parameters of the economy that determine the elasticity of the natural rate of interest with respect to the debt/GDP ratio, and on the growth rate of GDP. But we know it exists, and we know the US debt/GDP is increasing over time.

The UK has had 200% debt/GDP ratios several times, and has survived (though last time, post WW2, it survived largely through inflating away the debt/GDP ratio, and that is something that's a lot easier if popular memory still thinks in terms of gold standard price level dynamics, which it doesn't any more). But politically the US is not what the UK was. And the population growth rate was higher then. And Japan's debt/GDP is above 200% (though I can't remember if that's net or gross), and Japan's demographics don't look so good.

(4.) This isn't really about exchange rate crises. It would be easier to model what I'm talking about in a closed economy. And PK does seem to have a bit of a blind spot in thinking about the implications of debt in an OLG model.

Max: yep. I'm thinking more about inflation risk than default risk, because the US is not like Greece.

Min: Yes, today is very different from the 1980's. But if the bond vigilantes did attack, tomorrow would look a lot more like the 1980's.

Sergei: see my response to Simon.

(Yep, I confess I have gotten Max and Min muddled in the past. Crappy joke: I forgot to check second order conditions. Sorry.)

nathan: ?? because if it didn't, and if AD stayed too high, the price level would (eventually) explode. Standard macro model result, goes back (at least) to Wicksell.

Verner: you lost me. But the effects I'm talking about would be there in a closed economy too.

(Ugh....5:30 am here on the wet coast....)

"3. Empirically, the US budget was in deficit even before 2008, IIRC. A plausible assumption is that recovery would mean returning to that same primary deficit, plus the total deficit would be bigger because the debt/GDP ratio has increased since then. "

Disagree. (1) Bush tax cuts will have expired. (2) There have been further cuts in non-entitlement programs that will not automatically be restored as demand expands. (3) The wars are over and the soldiers are coming home. (4) Obamacare saves money. (5) Debt has been rolled over at lower rates. Where are the increases in the primary structural deficit that you think offset these factors?

In addition, if you're assuming intolerable levels of demand-driven inflation, aren't you thinking of cyclical pressures way beyond those of 2008? and therefore much better cyclically-unadjusted deficits?

I think that the problem with the Krugman analysis is that he has one interest rate. The Fed controls short term interest rates and this is probably what affects the dollar, but it is the long-term interest rate that affects domestic demand and it is long-term interest rates that have the "p" term. Risk goes up--U.S. long rates go up and the short rate and exchange rate are little changed (holding inflation constant as in Krugman).

"Bottom line: we don't know what the maximum sustainable debt/GDP ratio is... But we know it exists, and we know the US debt/GDP is increasing over time."

"To my mind, this reinforces the urgency for something like NGDP targeting. We don't want to wait for the bond vigilantes."

Okay, now I'm confused. I originally thought you were talking about something that you felt was urgent. But reading your latest, I get the feeling that this is something that might not happen for a generation.

I agree with you that PK doesn't worry much about debt-dynamics. To give the professor his due, he pretty frequently says that they have to be addressed in the vague-not-quite-yet medium term. So are you agreeing with him on that? Or are you trying to make the case that the US situation is "urgent"?

"This isn't really about exchange rate crises. "

Nick, if you want to make the above claim, then I'm more confused than ever. PK is explicitly talking about a model with exchange rate reactions and you are taking issue with his analysis. I don't understand why a speculative attack in a closed economy is considered to be so excessively expansionary that it generates inflation. I thought PK's model hinged on the exchange rate depreciation redistributing AD internationally. Could you please explain?

Must....sleep......more.....

Simon: demographics (aging boomers) would be one big effect going the other way. But sure, I would certainly claim no special knowledge about the US fiscal position. If you think that recovery would cause the US to swing back into overall budget surplus (with debt service costs adjusted for inflation plus long run real growth) then there's nothing to worry about. Yet. But the math tells us that can't be true forever if the debt/GDP ratio keeps on growing while we are waiting for the bond vigilantes.

ran: I see that as a complication that shouldn't affect the basic results.

Nick,

"A large attack by the bond vigilantes would be a bad thing, because it would increase Aggregate Demand too much. That would force the Fed to increase interest rates a lot, and that would force the US government to raise taxes and/or cut spending to cover the increased costs of servicing the debt."

It would force the government to change its financing and / or spending decisions, but you only listed two possibilities. Here are a couple of others:

1. Extend the duration of debt. Most government bonds are accrual type, meaning they don't make coupon payments. Instead they pay back interest and principle at maturity. And so, the federal government can escape the cash flow problem of higher interest rates by extending the length of time government bondholders must wait before they are repaid. Currently the average duration of the U. S. federal debt is something like 4-6 years.

1a. Paul is wrong if the federal government choses this route. If long term interest rates rise above the rate of inflation AND the federal government extends the average duration of its debt, then aggregate demand will fall. More than likely savings rates will rise.

2. Sell equity instead of debt. Government bonds (courtesy of the 14th amendment to the US Constitution) are guaranteed claims against future tax revenue. The federal government could just as easily sell non-guaranteed (equity) claims against the same tax revenue. This has important implications for productivity and the after tax cost of money in the private sector.

Max, you said:

"Different people have different conceptions of how government debt works. To some, government debt is a junior liability. Money is a senior liability. An excessive public debt (assuming most of the debt is bonds) then results in default risk but no inflation. No AD boost. This is I guess the more traditional view and what most people mean by "bond vigilantes"?"

I guess that would depend on your view of monetary policy. If you believe that monetary policy is truly independent from fiscal policy, then the currency is a liability of the monetary authority and government bonds are liabilities of the federal government. Federal debt is a senior liability of the federal government, meaning payments on it supercede all other government expenditures.

Long term government debt is often purchased by banks who borrow short to lend long (maturity transformation). That works as long as the short term interest rate set by the fed is less than the long term interest rate. Default risk can appear in bid to cover ratios at government bond auctions. Suppose an independent federal reserve (Not Bernanke and Fisher) decides enough is enough and pushes short term interest rates significantly above long term interest rates (inverted yield curve). Borrowing short and lending long is no longer profitable. This would affect the federal government's ability to rollover its existing debt (old bondholder is paid back by selling new bond).

"And then there's the fiscal theory of the price level / MMT view that regards money and bonds as equal government obligations, and so an unsustainable public debt causes a rise in the price level rather than an appearance of default risk."

Excessive public debt creates inflationary risk because the federal government can create a demand for goods (by borrowing money) without adding to the supply of goods. That increased government demand is offset by the monetary authority pushing interest rates on government debt above the rate of inflation.

Simon: PK has a Mundell-Fleming model, and models the attack as shifting the BP curve up, which increases AD if the exchange rate is flexible. You can ignore the BP curve, model the attack as an increase in expected inflation, stick a wedge between the IS and LM curves, and get exactly the same (qualitatively) increase in AD. It doesn't matter (for my point here) which way you model the attack - open or closed economy. What does matter for my point is how big that attack is, and how much AD increases, and how much r rises, and how that affects the overall budget deficit. And the magnitude of that effect depends on the debt/GDP ratio, which is growing over time while we wait for the attack.

If you see an enemy waiting to attack, and every year he grows bigger, but you think he won't attack until he's big enough to win, is that an urgent problem? It's certainly one you want to deal with sooner rather than later. Make him attack now, by NGDP targeting. Tease your enemy mercilessly. Fart in his general direction, don't make soothing noises. (Sorry, Holy Grail references).

Nick,

"Having long average maturity of the debt reduces rollover risk and gives the government more time to respond to a rise in r by raising taxes and cutting spending. So the risk of a sudden attack is lowered. But the government must still eventually respond."

Only if you believe that there is a limit to how long someone will lend the federal government money. Obviously that would be the case of an individual investor that has only a lifetime. But that would not be the case for institutional investors. I believe that Japan sells 50 year bonds. Would 100 year or 200 year bonds be out of the question?

This all comes down to how and when demand-pull inflation will materialize. If we're going to be serious about all of this, it seems we need to stop talking in generalities and actually model this stuff in a rigorous fashion. When, how, and why will X amount of spending cause X amount of inflation? What are the sensitivities and what do we need to believe for X to occur given how the economy may evolve? You need these details to truly understand the trade-offs and make good public policy decisions. To my understanding, no one is really doing this. There are people like Krugman saying spend now and cut later; there are people like you saying that this is dangerous; there are people saying people like Krugman and you don't really understand how govt spending impacts the economy. Is this not a huge gap in the research agenda?

In any case, the underlying tone is that the US/Japan wouldn't be able to cut spending or increase taxes enough to counteract stimulus from interest on the debt. I don't think that is a good assumption. We're already concerned about a debt crisis when the economy is nowhere near full capacity. Something tells me we'd act when there's an actual debt crisis.

Good post. This reinforces a point I keep making. Monetary stimulus is not "risky" as the Fed sometimes claims, rather slow NGDP post is highly risky.

Scott

"If you see an enemy waiting to attack, and every year he grows bigger, but you think he won't attack until he's big enough to win, is that an urgent problem? It's certainly one you want to deal with sooner rather than later."

Um.....no. (Seriously? That's all you've got?) Logic, please, Nick.

If I'm an elephant and my enemy is a mouse that is growing, I do not need to deal with the problem "sooner rather than later." Particularly if action in the short-term is likely to more costly than action in the medium-term.

The logic you're using here can be used just as easily to make the case for Massive Fiscal Contraction Now. If you think slow NGDP growth is highly risk (as Scott Sumner does, above), that's probably the last thing that you want to do.

Simon: "The logic you're using here can be used just as easily to make the case for Massive Fiscal Contraction Now."

No it can't. We want the bond vigilantes to attack now, not later. We don't want to do anything that would cause them to wait even longer. The whole point of NGDP targeting is to provoke an attack now. We want the bond vigilantes to say (and say it now, not later) "NGDP is going to rise, so it's time to dump bonds and buy real assets."

"It doesn't matter (for my point here) which way you model the attack - open or closed economy. What does matter for my point is how big that attack is, and how much AD increases, and how much r rises, and how that affects the overall budget deficit. And the magnitude of that effect depends on the debt/GDP ratio, which is growing over time while we wait for the attack."

Thanks, at least I think I understand more clearly the mechanism that you've got in mind. But I still don't get your point that "What does matter ... is how big that attack is..." Nick, you just explained a linear model, but I thought you were claiming a non-linear effect (small shocks good, big shocks bad.) Is this a debate about whether any AD shock causes the actual fiscal deficit to grow or shrink? Or have you got a non-linear Phillips Curve in mind (so that we don't get rapid expected deflation from a big output gap but too much AD would generate big expected inflation)?

"No it can't."

Read what you wrote. You argued that if you face a growing enemy, it should be dealt with sooner rather than later. That's the argument for We Must Balance The Budget Now. You clearly mean something else.

What I think you've done is confuse attack size and attack timing. Your original argument was about the size of the attack. Now you are arguing about the timing of the attack.

Simon: very briefly, because convocation is today:

Welfare is non-linerar in AD. Too little AD we get recession; too much AD we get hyperinflation. Absolutely standard assumption there.

But we want the enemy to attack. Because we want recovery, and recovery will both cause and be caused by the attack. Recovery and attack come together. But we don't want the attack to be so big it overwhelms us.

Size is a function of timing. The later the attack comes, the bigger it will be. I wish they had attacked sooner, but we can't do anything about that now. But we can get them to attack this year, which is better than next year, which is better than the year after.

We want an attack now, not later, not just because we want to end the recession now. But because a later attack would be bigger than an earlier attack.

BTW, I'm not sure there is anything here that PK would really disagree with.

"If you believe that monetary policy is truly independent from fiscal policy, then the currency is a liability of the monetary authority and government bonds are liabilities of the federal government."

Right. In this case, as long as the CB remains solvent, it can maintain price stability no matter what. People wouldn't flee from government bonds into real assets, they would flee from government bonds into money. This is of course what has happened in the euro zone.

If AD is increasing too fast, whatever else is true, you are no longer in a depression. The idea that it is possible to overshoot on a correction mechanism is no reason to absolutely reject that mechanism.

Nick,

I think the whole bond vigilante idea is a major cause of confusion. As you said, long term bonds are an unnecessary complication. So just think of t-bills. t-bills trade at the policy rate for straightforward arbitrage reasons. The policy rate is driven by a Taylor rule. So there are no bond vigilantes, neither in this model nor in the real world. Nominal rates rise when NGDP expectations rise as a direct consequence of some kind of forward looking Taylor type rule.

So the "bond vigilante attack" is *exactly* according to the equations that constitute our monetary policy. Nothing mysterious, nothing to fear.

The only question is, can our policy rule control inflation even in the face of high government debt? I don't see why not. What is it about high debt levels that makes demand *less* responsive to rate hikes?


@nick:aha! exactly what i thought. you're working with a natural rate of interest. I think in this context your thesis makes no sense. If you look at treasury yields, you'll see they are dominated by the target interest rate. If interest payments are driving inflation, a central bank that raised interest rates to slow inflation would be moronic. I understand how you get there ala wicksell, but i don't think that actually makes sense.

Nick, :

I find it in incredible that when both Canada and the United States are suffering a chronic lack of AD, you worry about too much AD. Who cares about a little inflation, the depression is worse.

Further to Nathan Tankus' point, in any real control system, you have and must have some degree of overshoot. A fairly realistic model (proportional/integral/derivative, second order differential) must have more overshoot the faster it works (gets to target first time). Then the system bounces around to settle down.

You argue that the Central Bank, which is exogenous here, its actions with regards to interest rates can be easily controlled, cannot restrain itself and will be overenthusiastic in stamping out a boom. What, central bankers can't be lazy and slothful? When the time calls for just that? We can't get Mark Carney drunk or stoned for a few months?

If the debt were small, the amount by which taxes would need to be raised and/or spending cut to cover the increased debt service costs would be small too, and it could easily be done. But if the debt were large, the amount by which taxes would need to be raised and/or spending cut to cover the increased debt service costs would be large too, and it could not easily be done.

This is one those statements about government debt financing that you put in these posts that contains more than one fallacy, IMO. It implies a static view of an economy, that an economy cannot outgrow its debt burden. That is exactly what the US and Canada did to our war debts. Tax rates remained static throughout the post-war period, in fact if anything they went down slightly, yet actual debt was never paid off in nominal terms. It was just paid off by out-growing it.

Your posts about taxes also posit that a government cannot borrow indefinitely from the private sector. As the UK's public debt is 300 years old, the US is 225 years old and Canada's has been around permanently since 1867, that supposition is wrong. With growth, which implies a multiplier greater than one, it is an open question whether debt is a burden. Most of the time it is not, as the interest rate on government debt is below the growth an inflation rate. Government debt is permanently ponzi-ish because the private sector allows it to be so.

Paul Krugman has repeatedly said, with good evidence, that the upper borrowing limit for a sovereign with its own central bank and currency is at least at 200 - 250%. The US ended WWII with a debt/GDP ratio of 120% and that was followed by the Post-war Boom. The UK government was a 200%+ and it still had decent economic performance. Japan is still not in trouble with a 200% ratio, so why we are worrying about the US is beyond me.

BTW if the bond vigilantes want to have a party in Canada, I'd be overjoyed. Throw them a party at the Bank of Canada, get them so drunk and stuff them with food they'd never want to leave. :)

The policy rate is driven by a Taylor rule. So there are no bond vigilantes, neither in this model nor in the real world.

Simple, clear and logical. I'm with K.


Nick, if you really believe there are such creatures as "vigilantes" you should make a general equilibrium case for that.

First, you need to throw away full Ricardian equivalence (otherwise the vigilantes would not care).

Second, assume a sudden change in preferences for less savings, causing everyone to go on a consumption binge as they attempt to sell their bonds and buy goods, even though the average family has less savings now than it did 10 years ago, and increased income uncertainty.

Third, assume that the central bank responds to this by hiking interest rates rather than tolerating an elevated level of inflation (which would reduce government debt). But even now, CBs all around the world are promising elevated inflation during the recovery.

Fourth, assume that the government is not receiving anymore revenue in response to the increased consumption binge, together with decreasing the benefit payments going out, allowing it to pay higher interest rates if necessary.


You have not painted a coherent picture of something that is likely to happen or that we should be afraid of if it does.

It seems to me that countries like Canada or the US can only be successfully attacked by bond vigilantes if its central banks joins in the attack, and if its government then passively accepts the joint attack and doesn't take legislative steps against it. In other words, these governments have to attack themselves.

Which is silly when the Government owns the Bank of Canada, has a memorandum of agreement about policy and in a zombie bond vigilante attack could issue a directive to the Bank of Canada to defend the CAD$, or buy a mountain of bonds, or whatever. It helps that the Bank of Canada is across the street and down a block from Parliament Hill so the Finance Minister can punch the light out of have a vigorous policy discussion with the Bank of Canada Governor whenever he wants.

Why havent they attacked already? You mean to tell me they have been happy with near 1% rate the last 4+ years? If there was any group actually able to demand higher rates, in the face of a fed committed to keeping the rates it controls low, we would have already seen them. We wont see higher rates til the fed says so.

The bond vigilantes are frolicking and doing blow with the confidence fairies.

Gizzard, I don't think Nick is saying they'll force higher rates. He's going along with Krugman and saying an "attack" will manifest through exchange rate depreciation.

wh10: He's going along with Krugman and saying an "attack" will manifest through exchange rate depreciation

Hm. How do you attack a floating exchange rate? I mean attack permanently. Floating exchange rate is a consequence, not a reason.

BoJ has been attacking exchange rate for ages and despite its unlimited firepower those "vigilantes" seem not to care and keep on pushing.

Simon:

"I think that the problem with the Krugman analysis is that he has one interest rate."

It goes beyond that. The problem with the Krugman analysis is that it does not distinguish between accrual debt that the federal government sells, amortization debt that the private sector engages in, and coupon debt that the federal government used to sell.

The only way that interest rates "control inflation" is by making a transfer from positive cash flow to positive balance sheet. Meaning that by buying long term accrual government bonds a person is willing to forgo positive income for positive net worth.

Picture a federal government that only sold coupon bonds and a private sector that had an excess of aggregate demand.
Question: How would higher interest rates lower aggregate demand?
Answer: They would not, in fact they could make it worse. A rise in interest rates would mean an increase in coupon payments by the federal government.

Nathan,

"That is exactly what the US and Canada did to our war debts."

WRONG. Here is a chart comparing total US debt (public and private) to US nominal economic growth:

http://research.stlouisfed.org/fred2/graph/fredgraph.pdf?&chart_type=line&graph_id=&category_id=&recession_bars=On&width=630&height=378&bgcolor=%23b3cde7&graph_bgcolor=%23ffffff&txtcolor=%23000000&ts=8&preserve_ratio=true&fo=ve&id=GDP_TCMDO&transformation=lin_lin&scale=Left&range=Custom&cosd=1950-01-01&coed=2012-07-01&line_color=%230000ff&link_values=&mark_type=NONE&mw=4&line_style=Solid&lw=1&vintage_date=2012-11-11_2012-11-11&revision_date=2012-11-11_2012-11-11&mma=0&nd=_&ost=&oet=&fml=a%2Fb&fq=Quarterly&fam=avg&fgst=lin

The war debt was not reduced by growth. It was replaced with private debt. There are only two ways to reduce debt. You either collect more in income than you spend - which is impossible for everyone to do in a capitalist system. Or you convert debt to equity.

Frank: "There are only two ways to reduce debt. You either collect more in income than you spend - which is impossible for everyone to do in a capitalist system. Or you convert debt to equity."

Is this what you learned from MMT??? No, no, no! If this year I spend $100 more than my income, and you spend $100 less than your income, and lend me $100, then debt rises by $100. Next year I spend $105 less than my income, you spend $105 more than your income, as I repay my loan and interest, then debt falls by $105.

But that is totally off-topic.

I read the comments, then despaired, and went and changed the oil and filter on the MX6. rsj is sort of maybe getting it (though writing down the equations of a model and solving it is a lot easier said than done). But the rest seem to be totally misunderstanding my post. Now clearly that's maybe at least partly my fault, for not explaining myself clearly enough. Or maybe it's just a hard topic to explain clearly. Or maybe you're all coming from some totally different MMT-space. (Dear old Winterspeak totally missed my point as well).

Aggregate Demand depends (negatively) on: the rate of interest; the willingness of people to hold the existing stock of government bonds at a given rate of interest rather than sell them for money and then use the money to buy real goods.

AD is too low in the US right now. The US needs (say) a 5% increase in AD. But you don't want AD to increase more than that. There is an AS curve out there somewhere! We don't want an infinitely large increase in AD.

If people became less willing to hold US bonds, that would increase AD. But if it increased AD by too much, the Fed would have to raise interest rates to stop AD increasing too much. If people became a lot less willing to hold bonds, the Fed would have to increase interest rates a lot.

I think I will now wash the MX6.

If people became less willing to hold US bonds, that would increase AD. But if it increased AD by too much, the Fed would have to raise interest rates to stop AD increasing too much. If people became a lot less willing to hold bonds, the Fed would have to increase interest rates a lot.

Overshoot is bound to happen in any realistic system with inefficiencies, constraints and storage dynamics. Tell the Bank of Canada Governor to take up pot smoking while this recovery is taking place, so he doesn't get all antsy about inflation.

The faster the rise time (from initial condition to target), the greater the overshoot. That's math.

Why are you seeing inflation under every bed, Nick? Real work requires the worker to get his hands dirty, so a little inflation is no bad thing.

Nick,

First, worrying about excessively high AD is like shouting fire in Noah's flood.

Second, if there was a sudden and dramatic shift for less risk free savings -- something that in the history of the earth has never occurred -- all that the CB would need to do would be to raise rates by a small amount, not by a large amount.

This is because of duration. Right now, rates are extremely low -- yield on a 10 year bond is 1.6%. Suppose all government bonds were 10 years. Suppose there was a sudden demand to decrease holdings of government bonds by 10%. How much would the CB need to raise rates to accomplish this? About 1.2%.

In a low interest rate environment, the CB would only need to raise rates by a very small amount in order to accomplish a large reduction in savings. That is why Japan has been stuck at zero rates for about 20 years now. If this was really an "unstable" system, we would have seen it. But when you have long lived assets in a low rate environment, very small changes in rates correspond to large reduction in savings.

I don't think you are looking at this properly and I don't think you have a convincing model.

But again, it's all shouting fire in Noah's flood.

A better metaphor: I'm saying we need to fix the leak in the reservoir as soon as possible, because when the flood is over we won't have enough water left for the dry season.

Nick,

"There are only two ways to reduce debt. You either collect more in income than you spend - which is impossible for everyone to do at the same time in a capitalist system. Or you convert debt to equity."

I typed too fast - missed the "at the same time" part. I think we can agree that simultaneous surpluses are impossible.

"Is this what you learned from MMT??? No, no, no! If this year I spend $100 more than my income, and you spend $100 less than your income, and lend me $100, then debt rises by $100. Next year I spend $105 less than my income, you spend $105 more than your income, as I repay my loan and interest, then debt falls by $105."

Again I missed the "at the same time" part. Sorry for the confusion. As for MMT, it has a lot to offer in terms of economic theory, but even it does not consider the possibility of government equity. MMT explains the federal government's lack of bankruptcy risk (government bonds denominated in government currency) but it misses a few components - the yield curve matters and the difference between public and private debt (amortization versus accrual) matters.

You said: "A large attack by the bond vigilantes would be a bad thing, because it would increase Aggregate Demand too much. That would force the Fed to increase interest rates a lot, and that would force the US government to raise taxes and/or cut spending to cover the increased costs of servicing the debt."

A large attack by the bond vigilantes would neither be a "good" or "bad" thing. Apparently you believe that economics is some sort of morality play with good guys and bad guys. It wouldn't force the US government to do anything. To accomodate the increase in the cost of servicing its debt the federal government could chose to either increase revenue, cut spending, increase the duration of its debt, or sell equity instead of debt.

You said: "The longer the US stays in recession, with a large budget deficit, the bigger the debt will grow."

The longer economists and economic advisers keep their models limited to a Hicksian three good economy (money, bonds, and commodity goods), the bigger the total debt (private + public) will grow.

When water is gushing over the top of the reservoir dam, a leak is the least of your worries.

You cannot simultaneously try to fight inadequate AD and some imagined, theoretical inflation crisis. First things first.

Why exactly do we need to invent non-existent problems when our actual problems are more than troublesome enough?

Aggregate Demand depends (negatively) on: the rate of interest; the willingness of people to hold the existing stock of government bonds at a given rate of interest rather than sell them for money and then use the money to buy real goods.

I don't get it. If one guy sells his bonds for more money, then he has fewer bonds and more money; but some other guys has more bonds and less money. How did this increase aggregate demand?

Dan Kervick,

"I don't get it. If one guy sells his bonds for more money, then he has fewer bonds and more money; but some other guys has more bonds and less money. How did this increase aggregate demand?"

Great question. For that you have to understand the yield curve and the nature of government debt. The correlation between aggregate demand and short term interest rates is positive, not negative. The correlation between long term interest rates and aggregate demand is negative. And so what Paul Krugman and Nick Rowe are really asking for is an inverted yield curve. Low long term rates (to discourage saving) and high short term rates (to increase income and aggregate demand).

Dan: suppose there were a rush to get out of bonds to hold money instead. No problem. The central bank just takes the other side of the trade, and buys the bonds (unless we are talking about Greece, of course). But if there were a rush to get out of bonds and into real goods, via money, then a little bit of that is a good thing, but too much of that increases demand for real goods too much. They sell the bonds for money, *because everything gets sold for money*, but they don't want to *hold* the money, they want to spend it.

Dan,

"But if there were a rush to get out of bonds and into real goods, via money, then a little bit of that is a good thing, but too much of that increases demand for real goods too much."

Which is a big IF. People have an amount of investment / savings demand and so if they sell their bonds they may not buy more goods but may instead buy some other financial instrument (bonds of another country, equities, etc.). That is why they bought the bonds in the first place, which Nick seems to forget.

"They sell the bonds for money, *because everything gets sold for money*, but they don't want to *hold* the money, they want to spend it."

Or they want to buy something that has a higher potential rate of return.

Dan,
You asked the question: "How did this increase aggregate demand?"
Answer: It doesn't, at least not in any direct fashion. Ultimately any increase in aggregate demand is a choice. Milton Friedman believed that people make their spending decisions based upon anticipated future income instead of current income. Lower interest rates reduce that future anticipated income.

Nick Rowe said:

"PK does seem to have a bit of a blind spot in thinking about the implications of debt in an OLG model"

Blind spot??!!

Ridiculous. He understands your ideas about the issue, but rejects them.

Don't imply you're smarter than "PK" and that he can't "see" what you do. You're certainly not and he certainly can.


Again on the reservoir metaphor:

If a flooding reservoir's dam has a hole in it, that is no bad thing. The flood made the hole, and we have to get rid of the flood before we can repair the hole. We can always sink a cofferdam when the flood is gone, a quick and easy procedure with that spare cofferdam we have on hand, and then make a thorough repair. But we have to get rid of the flood first before our dam can be save to work on.

I think I'm lost, Nick.

First of all, if the rate of interest changes, how does that affect people's behavior with respect to the existing stock of bonds, whose interest rates don't change? But more importantly, suppose for whatever reason the holders of the existing stock of bonds develop an increased aggregate desire to sell them. They can only succeed when there are other people willing to buy them at the offered prices, which will be driven lower. When these transactions occur, some quantity of money changes hands. But the total amount of money doesn't increase - it just belongs to different people. So how does this increase aggregate demand? Are you assuming these sales represent a net shift from people with higher propensities to save to people with lower propensities to save?

Anyway, I don't see what all this business about the existing stock of bonds has to do with an "attack" of the bond vigilantes, which I understood in the past to refer to the possibility of the bond market forcing falling prices/higher yields on new debt. Krugman is arguing that if the risk premium for government bonds increases, then since the Fed sets the interest rate, the result would be a fall in the currency, which would increase aggregate demand. And you are saying, I gather, that if the increase is too big, the Fed would then be forced to let interest rates rise to decrease aggregate demand. So I guess you are just disagreeing with Krugman's model?

Anyway, I don't get this business about the risk premium - I guess in neither Krugman's model or your thinking. Unless the government stupidly threatens to voluntarily default, there is never any reason for rational bond investors to worry that governments like the US and Canada won't redeem their bonds. They can't run out of money.

Nick: you seem to be using odd terminology. I think you're saying that the propensity to spend out of bond wealth has increased. Am I correct?

Nick

I'll just repeat the comment I left at Tyler's post at MR:

” In essence, in his model, the central bank can push down “interest rates” (in general) without bearing any inflation repercussions”

This really is the key, crux of the entire debate. What latitude does the Fed have, should it need to, in imposing -ve real rates at the long end of the term structure and getting away with it. For how long? US long bonds are *money*, globally speaking. As a cherry on top is the -ve beta hedge nature of any long govvie bonds (something I borrow from commenter K). Combined, this means that the Fed has significant latitude in imposing -ve rates at the long tenor.

So, I think Krugman will win the *empirics* of this, though not for the reasons he mentions. His model is woefully under-developed and was rightly taken apart by Tyler.

I don't share K's rule-based monetary policy enthusiasm, by the way. The (im)potence of monetary policy cuts both ways - we concrete steppe Wicksellians can't have our cake and eat it too. In developed economies, currently, central bankers seem unable to raise inflation by Chuck Norris-ian fiat and their bazookas seem weak. In many other economies, indeed in the vast majority of economies around the world, central bankers have struggled to win the fight against inflation, by fiat or by bazookas. Indeed, historically, bazookas meant for *tightening* policy have been weaker than those meant for stimulating the economy.

Yes, a determined central bank can definitely get inflation under control by imposing sufficiently high real rates, Taylor rule or no rule. To do this, it needs to 1) be willing to crush private demand for an unknown, even if short, duration 2) be willing to impose greater interest payout burden on the government in the short run.

1 is what Volcker did. Indeed, it is what many emerging market central bankers have also done when faced with disturbing inflation dynamics - e.g. Rangarajan & RBI in India '96.

2 becomes more severe, and hence more unlikely, the higher the level of government debt is. It is not the size of the debt itself that causes the inflation, it is the sensitivity effect of large sovereign debt on the central bank's willingness to raise real rates that is key.

Christopher Sims makes the exact same point in many of his papers - the inflation fighting stance of central banks is compromised in a regime of high sovereign debt.

In fact, as I tried to argue over at Cullen Roche's a few weeks ago, the central bank's ability to backstop any sovereign bond auction is a double-edged sword. Bond defaults won't happen, hyperinflaion won't happen. But they're both red herrings and fiscal conservatives who talk up those things are either being Straussian or wasting their time. The real non-trivial threat is high and rising inflation, which is much more common than we like to admit these days. And the size of public debt has something, even if indirectly, to do with the monetary authority's ability to combat high inflation should the need ever arise.

As a last, somewhat un-related point, the empirics of all this are severely challenged due to Milton Friedman's thermostat. The most probable event is that not much would happen - the US will manage to avoid high inflation. How would we know whether this was because the bond vigilantes didn't come, or whether this was because the bond vigilantes were rendered unsuccessful by the Fed? The data would look almost similar either way.

Ritwik,

"In developed economies, currently, central bankers seem unable to raise inflation by Chuck Norris-ian fiat"

Can you give an example?

W.Peden

Japan, Switzerland, US (partially). I know, for all of those one might say, they haven't done this, that or the other, which would have ensured they hit any target they wanted. Those would be fair points, though I'd disagree. But I don't particularly want to get into that debate on this thread. On this one, I'm running for Wicksell from the right flank instead of the left one. :)

Dan: there are 3 risks to holding govt bonds: outright default (unlikely for a govt that can print); inflation reducing the real value; (in an international context) an exchange rate depreciation. (PK's way of modelling the attack is equivalent to assuming an increased risk of exchange rate depreciation). To my mind though, what matters most in the current US/Japan context is the risk on bonds *relative to* the risk on real assets. It's not that people might become more scared of holding bonds; they might become less scared of holding real assets. If people became more confident of recovery, they would be less scared of holding real assets and so less willing to hold bonds instead of real assets. The fact that people are willing to hold US govt bonds at (mostly) negative real interest rates is very abnormal. It's both a consequence and a cause of the recession. It won't last.

"First of all, if the rate of interest changes, how does that affect people's behavior with respect to the existing stock of bonds, whose interest rates don't change?"

Terminology. If interest rates rise, the market prices of existing bonds will fall, and so the *yield* on existing bonds will rise. Most macroeconomists use "yield" and "rate of interest" synonymously.

"But more importantly, suppose for whatever reason the holders of the existing stock of bonds develop an increased aggregate desire to sell them. They can only succeed when there are other people willing to buy them at the offered prices, which will be driven lower. When these transactions occur, some quantity of money changes hands. But the total amount of money doesn't increase - it just belongs to different people. So how does this increase aggregate demand?"

You are (implicitly) assuming the velocity of circulation of that money is fixed. If both M and V were fixed, then yes, AD would be fixed too, and nothing else would affect AD. Even I'm not that extreme a monetarist! When people try to sell bonds, bond prices fall, which means the yield (interest rate) on bonds rises, which increases the opportunity cost of holding money, so velocity increases, so AD increases for a given stock of M. And if the central bank were to increase M to prevent interest rates rising, that would cause AD to increase for a given V. So either way MV increases and AD increases (unless the central bank were to reduce M to offset the rise in V to prevent AD increasing).

" And you are saying, I gather, that if the increase is too big, the Fed would then be forced to let interest rates rise to decrease aggregate demand. So I guess you are just disagreeing with Krugman's model?"

No. If you look at PK's model, he has the Fed set r as a positive function of Y. If AD increases, and Y increases, PK assumes the Fed will increase r. (I would prefer to modify his model by assuming the Fed will hold r at 0% until Y increases to "full employment Y*" and then increase r by whatever it takes to prevent AD rising any further past Y*.)

Nathan: it's not me, but you, using odd terminology ;-) You are using Old Hydraulic Keynesian terminology. But describing it as an increased marginal propensity to spend out of bond wealth would mean basically the same thing, as long as you allow that that marginal propensity to spend also depends (negatively) on the rate of interest.

RN: Yes, I am denying the Sacred Doctrine of Pauline Infallibility, even though I recognise (obviously) that PK is a better economist than me. Nobody is infallible.

Nobody has done as much damage to the doctrine of Pauline Infallibility as PK himself.

Frank Restly. CApitalist and money-using are so not the same thing. Not only a pedantry point but we need to be very precise in our definitions.

Nick: We are not thinking clearly. That's why we debate. Though it helps when we use the right words.
I will not argue about "burden of debt " or such. Not because it is not really the precise subject of this thread but because of the sage advice of a british prime minister : " Three persons understand the problems of the Balkanns. One just died, the other went mad and as for me, I don't want to talk about it." And a Governor of the Bank of England later piled on;" Two people understand the working of the gold standard. Me and the Chief Cashier. We disaagree."

@Nick: fair enough :P it does seem to me odd not to explain it in terms of consuming out of wealth because otherwise people think you're crazy for going on a spending binge because you got a different form of your old asset. Do you usually go hog wild when your saving account matures?

Nick,

Lets assume a closed economy with all government financing in t-bills. Now obviously the "bond vigilantes" can't attack via selling debt and directly causing a rise in yields. Yields are directly determined via arbitrage vs the policy rate. What people *could* do is borrow in order to spend or invest, which will raise inflation, leading to higher rates via a Taylor rule mechanism. But first you have to explain *why* they would do that in response to high government debt. I.e. you need to tell us why the equilibrium real short rate required to stabilize inflation increases with the level of government debt. I think there may be good explanations, but I don't think it's obvious (see Japan) and I don't think you've connected the dots.

I have a distinctly different view of the situation in the USA in which the attack of the bond vigilantes is a pretty bad thing. For me it looks like the US have found an arrangement that comes close to the government buying goods with newly printed money as long as it does not push up inflation or the trust in government bonds/money too much. In this case the attack of the bond vigilantes would be a bad thing as it would make everyone poorer:
http://makrointelligenzint.blogspot.de/

Ritwik: I think (from Tyler's thread) PK has a legitimate defence for "changing his mind" in this case: PK2007 assumed full employment; PK2012 assumes unemployment. The facts changed, so he changed his mind/model. And what I'm doing is saying "Hang on, if the drop in demand for bonds is big enough, we will go from the unemployment model to the full employment model, and we need to look at both."

K: If a "bond-vigilante attack" means a desired switch from bonds *into US money*, you are right. The Fed just buys bonds for money, and keeps r constant (to prevent r rising and AD falling). But if there's a desired switch out of bonds *into anything else* (foreign bonds, as in PK's model, or directly into real goods) then AD will rise unless the Fed raises the policy rate to make holding bonds more desirable.

nathan: sorry, you lost me there.

Jacques Rene: I thought that was the Schleswig-Holstein question, and Bismark?? (I could be wrong).

suppose we have a world in which the government prints so much money and buys goods with it until inflation is stable, do you agree with me that in this case a loss of confidence of international investors in government bonds would make the domestic people poorer and lead to higher unemployment, Nick?

Nick,

"But if there's a desired switch out of bonds... into real goods"

Yes, but *why* would higher government debt imply a greater desire for goods over savings at the *same* real rate? You still haven't told me *why* the equilibrium real rate increases as a function of the quantity of government debt.

The way I see it, high government debt could *depress* the natural rate by raising expected future taxes and thereby disincentivizing investment. I'm not saying you're wrong - I just don't see where you've explained relationship between the natural rate and government debt.

K,

Assume bond default is impossible. If the government is perceived to be insolvent, the price level will rise to make the government solvent (by reducing the real value of the public debt). Note that the central bank can't prevent this inflation, because money and bonds are part of a unified public debt. The CB can only control the price level independently of fiscal policy if bond default is possible.

Of course, solvency is not an objective reality. It's a self fulfilling perception. Very unstable. It's hard to be just a little bit insolvent.

higher government debt -> higher expected inflation because of a fear of a sudden run

Nick

On PK, I don't refer to his changing his mind, which is something I'd like to see more people doing! In general, over the past few years, he has proved to be an incredibly bad parser of developing macroeconomic thought, both mainstream as well as heterodox. More and more, he reminds me of Bob Solow, who managed to quip seni-funny witticisms on both Bob Lucas and Hy Minsky, without ever grappling truly with either of their critiques. His toy models have gotten more and more toyish, to the point that hey are no longer useful now.

Like Tyler said, the question should be considered as a theoretical question conditional on the assumption that the bond vigilantes do arrive. PK was claiming they don't matter even when they arrive, as a matter of pure theory. I'm saying as a matter of empirics it probably won't, but the theory is not remotely as clear-cut as PK implies it is.

K

1) I don't know about levels of debt but surely the delta of government debt matters for the natural rate? Pushing the IS curve up or down etc. How else would you analyse fiscal policy in Wicksellian terms.

2) Even if the economy is assumed to be in debt/GDP equilibrium, I think that a higher equilibrium level of debt makes the central bank less likely to correct Wicksellian errors due to the short term impact of such correction on interest payments.

K

"The way I see it, high government debt could *depress* the natural rate by raising expected future taxes and thereby disincentivizing investment"

Tobin 1978 has a very good analysis of this, where he basically concludes that steady-state high government debt/GDP *levels* can either crowd out private investment, or cause inflation, but not do both at the same time.

http://cowles.econ.yale.edu/P/cd/d05a/d0502.pdf

My take is, the balance of probability lies in favour of inflation, given that at low real rates, the corporate sector can produce many kinds of spending demands even without business investment, including forward-buying of commodities. Diego Espinosa likes to make this point about inflation in Lat Am. Again, the government debt acts primarily through the effect it has on central bank actions - a CB faced with high debt and the prospect of rising sovereign interest payments is more likely to *let* low real rates persist at the the long tenpor for too long.

Nick, Max, Ritwik,

If the representative investor is holding more t-bills, and less private capital (as a result of future taxation), then they are going to demand a higher rate of return on their t-bills relative to stocks. In other words, the equity risk premium declines the greater the share of investible assets is made up of government bonds. But at the same time, growth will be negatively impacted by expected future taxation which will depress yields on all capital asset, private or public. So we have two offsetting effects.

I think the first effect is very small for two reasons:

1) Inflation targeted t-bills get a very small risk weighting in investor portfolios so long as the government is expected to be able to deliver on its inflation target.

2) To the extent that US government bonds make up less than 20% of investible US assets a) they are not yet a major claim on all future GDP and therefore there is not a significant threat to the future ability of the government to make good on them via taxation (rather than inflation) and b) even if they were considered a bit risky, they still make up a very small proportion of total investor risk and we can therefore still expect a very large equity risk premium: t-bills, as an asset class, are not significantly accretive to the risk of the market portfolio.

The second effect, the depression of the natural rate via disincentives to investment from future taxation, is likely to be *huge* well before the outstanding quantity of government bonds gets anywhere close to half of all capital assets, and certainly way before government bonds make up half the risk.

Just to clarify my last comment: When I say that investors will demand a higher return on t-bills compared to stocks, what that means is that since the CB has fixed the return on t-bills, stocks will rally which thereby will result in additional investment which raises inflation. But, the equity risk premium will only decline to zero once perceived risk on all t-bills is equal to the perceived risk on all other risk assets. That will require a *very* large level of government debt.

Ritwik,

I had not seen your last comment yet when I wrote my previous comment. I will check out the Tobin paper.

K

I'm not sure why you're requiring that the equity risk premium decline substantially for the government debt to truly matter. High inflation scenarios are often the result of sustained spending by those private actors who have access to near-govt rates of cost of capital. Private debt financing is more relevant here. Somewhere, that spending has to set off a spiral of input prices - wages and commodities.

Ritwik,

I don't think it's even a defensible position that Switzerland, the US and Japan HAVE inflation targets and are missing those targets. In the case of Switzerland in particular, it's very uncontroversially false: an economy with free movement of capital can have an exchange rate target or an inflation target, but not both. If you mean "the Bank of Switzerland has had to engage in open-market operations to defend its exchange rate, which it has successfully done thus far despite doubts from some people", then that is true- but how much does the Bank of Switzerland's SUCCESS tell us about the impotence of monetary policy? I would argue: very little.

In the case of the US, there is also no explicit inflation target, but 2% CPI inflation in usually regarded as the rough number that the Fed likes to approximate. What was the US CPI inflation rate in September? 1.99%. Perhaps a shortfall of 0.01% from an unofficial and opaque target tells us a lot about the limits of monetary policy, but I have my doubts. Of course, one might assert that this is a fluke month, but then one looks at CPI inflation in 2011: 3.16%. Quite a performance at the ZIRP. Is that what is meant by a central bank being "unable to raise inflation" by Chuck Norris fiat?

That leaves Japan, which has been targeting inflation for less than a year. There, the explanation may well lie in what they have (or have not) done since February.

So I really don't see the phenomenon you apparentely do (central banks unable to create inflation) in the actual data. Then again, I'm not as far from the concrete steppes as some people, and your emphasis might have been on the Chuck Norris side of things. My preference is to think in "Chuck Norris" terms, but to also remember that people first have to know that Chuck Norris is willing and able to kick.

W,

If you observe a gas pedal pegged to maximum, you can infer that the driver is going slower than he wants.

Central banks have been "flooring it" (bank rate at minimum) since 2008 so it's reasonable to conclude they are missing their targets.

This is also consistent with the positive stock market reaction to inflation, which is abnormal. Normally inflation makes people worry about monetary tightening and recession.

Ritwik,

"I don't know about levels of debt but surely the delta of government debt matters for the natural rate? Pushing the IS curve up or down etc."

Maybe, maybe not. The Ricardian/Wallace view is "not." I dunno. You tell me.


"I'm not sure why you're requiring that the equity risk premium decline substantially for the government debt to truly matter. High inflation scenarios are often the result of sustained spending by those private actors who have access to near-govt rates of cost of capital."

Exactly. So we need to figure out if higher government debt causes higher spending/investment by private actors for any given real rate, right? That's what I am doubting. I think raised tax expectations reduce the desire for investment and current consumption. That *decreases* the natural rate.

So what could *raise* the natural rate. The way most people want to think of it is that if you increase the supply of something, the price has to decline. When you are talking about capital assets, that's a portfolio balance effect. The problem with t-bills is, to the extent they are risk-free, they don't have any risk weight in a traditional portfolio choice model. (Also, they aren't a *net* asset in the economy: for every t-bill asset there's a future tax liability. That's the Ricardian perspective.) But lets assume that an increasing quantity of government debt increases the inflation *risk* associated with t-bills. I.e. it doesn't increase the expected amount of inflation, but maybe it increases the outlook for inflation volatility. Maybe inflation is hard to control if there is lots of government debt. That would make t-bills less desirable than previously, thereby compressing the equity risk premium.

Lets assume that the market portfolio yield that equilibrates desired current and future consumption is unchanged. That means that at equilibrium, if the equity risk premium declines, stocks have to yield less (rally) and t-bills have to yield more. I'm betting that the tax disincentive effect is much bigger than the portfolio balance effect and overall, equities don't rally on surprise increases in government debt.

Nick: I wrote fast between two course. It was Schleswig-Holstein, Though there is a link. Bismarck said that half the problems of the world came from the abuse of alcohol and the other half from the Balkans.
I still won't enter this thread...

Max,

I don't accept the gas pedal-interest rate analogy, since a 0% bank rate is consistent with a huge variety of possible inflation outcomes- including 1.99% inflation in the US. Given the size of asset sheets, the quantity of excess reserves and market expectations regarding QE in many developed countries, it is perfectly possible for central banks to bring down inflation without raising bank rate. Hence the Bank of England has gradually brought CPI inflation down to its target level over the past few years without having to raise interest rates.

Of course, if you had a country that made an unlimited monetary base commitment to a nominal target (like the Swiss Franc-Euro peg) and failed to meet it, the gas pedal analogy would definitely be relevant. However, we don't have to play detectives to work out targets for most central banks. If you assume a 2% target for the US, the Fed has (insignificantly) OVERSHOT its target from 2008 onwards. The Bank of England has only had sub-target inflation for a few months in the past 10 years.

Sometimes reading central bank objectives is like reading an Agatha Christie novel. For example, Japan's price level target could only be inferred after years of seemingly incoherent policy (big expansions in the monetary base followed by tightenings when deflation eased off) started to form a clear deliberate pattern. Today, though, we have both explicit or widely implied central bank mandates for most developing countries AND their internal forecasts. We can infer that the Bank of England doesn't really mean 2% over the medium term when it says 2% over the medium term, thereby salvaging something from the theoretical wreckage of the Keynesian (as opposed to Krugmanite) liquidity trap and saving a gap in macroeconomics for fiscal stimulus, but I don't see the facts as being consistent with the arguments.

Nick, you said:

"There are 3 risks to holding govt bonds: outright default (unlikely for a govt that can print); inflation reducing the real value; (in an international context) an exchange rate depreciation. (PK's way of modelling the attack is equivalent to assuming an increased risk of exchange rate depreciation)."

There is actually a fourth risk. Suppose you buy long dated bonds with an anticipation that the federal government will pay you your principle plus accrued interest at the time the bond reaches maturity. Suppose after some time period the federal government decides to run a surplus and pay down its outstanding debt on an accelerated basis (prepayment risk).

I realize that it does not happen very often, but it does become a risk for pension funds and other insurance type enterprises that must try to match the durations of their liabilities with the durations of their assets.

K

"Exactly. So we need to figure out if higher government debt causes higher spending/investment by private actors for any given real rate, right?"

Not really. The mechanism in my head is:

1 High pvt spending and rising prices is the result of the real rate being kept too low by the central bank.
2) The central bank keeps the real rate too low because
a. It is incompetent and/or compromised
b. It isn't generically compromised, but is weakened in its resolve as it doesn't want to force sharp increases in government interest payouts.

2b is the relevant scenario for the US, I think. It's saving grace is that 'too low' real rate required to set off the spiral presently seems to be so low that the Fed can't seem to reach it, and even if it does manage to reach it, it has considerable latitude in making an error due to dollar being reserve currency, long dollar bonds being the ultimate risk-off asset etc.

Which is why I said that Krugman will probably be right about the empirics, but not because his model is solid.

Ritwik,

I don't agree. What you are talking about is a very short term CB failure. If the CB holds the policy rate too low, inflation won't just be too high - it'll accelerate upwards and it won't stop until they hike rates by more than the excess inflation. Then they'll bring it back to the desired 2%. All which will take a few years and be a painful reminder of 1980 in case any central banker has forgotten it (which I seriously doubt). And after that pointless escapade we'll *still* have to answer the question of what has happened to the natural rate as a result of increased government debt. Which is the interesting question that we are trying to answer here. 

Frank Restly,

Government bonds are not callable. If they want to buy them back, they have to pay the market price, just like any other buyer who wants to buy up the entire issue. If you don't want to sell, they can't have them. So there's no prepayment risk.

K

1. I don't think the Ricardian model holds. Gov't borrowing and pvt borrowing are not substitutes. Given a certain amount of leverage in the economy, government borrowing helps the pvt economy unlever and pushes up the discount rate at which the representative investment is viable. MMs like to analyse it as a function of higher NGDP expectations (cash flow numerators), monetary policy, etc. but I go along with the Keynesian version that says the mental discount rate is the key.

I'm relatively sure that this holds for delta(debt). I'm not sure I can make the same argument for a higher *level* of steady state government debt.

2. Even if the Ricardian model holds, it only claims that debt increase doesn't push the IS curve upwards because the expected tax increase brings it back. It would be hard to claim that the natural rate gets depressed as a result. At best you could say that it is unaffected.

A higher level of steady state government debt is important for more robustness/sensitivity/game-theoretic response-of-CB reasons than for what it does to the natural rate. At least in my mental model.

Also, if the *short term* CB failure lasts a few years, then that's pretty much all there is to it! There's hardly a macroeconomic crisis that lasts more than that.

W Peden

re Switzerland, you're mis-stating the trilemma. The Swiss peg only means that they lose control of the monetary base. It doesn't mean that they necessarily lose control of domestic inflation, which has been between -1% and 0% for a long time. Unless your argument is that the Franc is still overvalued at 1.2 to the Euro, and so the SNB is *passively* tightening even at that level.

In the US, GDP deflator was up 1.1% in 2009, 1.2% in 2010, 2.7% in 2011. Yes, it's not a particularly deflationary scenario. But I'd be hard pressed to argue that the '11 pick up happened on account of the Fed's fiat.

Japan has been targeting inflation for less than a year, but do you really think they were ok with -1% inflation when they weren't targeting inflation?

Further, none of these scenarios need to be construed as binary full control vs no control ones. It seems roughly true to me that the CBs in none of these countries have as much *traction* as they'd like over AD. The debate over whether this is self-induced or external can last ad nauseam, so I'll skip that here.

K,

In the US that has been the case since 1985, although there are still callable Treasury bonds out there. See:

http://www.ehow.com/info_7796395_treasury-bonds-callable.html

Frank,

If the US is still paying 1985 vintage coupons on par-callable bonds then the US government is an idiot.

Ritwik,

"Japan has been targeting inflation for less than a year, but do you really think they were ok with -1% inflation when they weren't targeting inflation?"

Yes, although they were happier when it was closer to the -0.5% to 0% range. I do know that they preferred it to anything much over 0%.

In the case of Switzerland, the Swiss central bank is responding to the strongest pressure, which is on exports. I think they're pretty happy with a stable exchange rate and mild deflation; getting the latter over the 0% mark would require modifying the exchange rate peg, which would defeat the stability objective. So it is an impossible trinity situation, though different in some respects from a peg. It's a floor, not a peg, because the Swiss central bank does not have to act in response to a weakening of the Franc as in recent weeks and has not reacted to keep the rate close to 1.2 in response to this minor devaluing.

It's not that the Swiss Franc is overvalued against the Euro at 1.2 (I'm really not familiar enough with international currency theory to say) but rather that, as with an interest rate target, they'd have to move the floor around in order to hit an inflation target. Given that the target was sold on the basis of stability for exporters and importers, I think that's unlikely. It shows one of the limitations of exchange rate targeting, at least when its justification is based on exchange rate stability.

The Fed does not focus on the GDP deflator and its 2% implicit target is CPI inflation, which has generally been faster than increases in the GDP deflator. GDP deflator inflation was about 0.5% slower from 1992-2008, implying a 1.5% GDP deflator target in comparison to a 2% CPI target, so a hypothetical GDP deflator implicit target would have been very slightly overshot in the years you mention. All this has been compatible with the Fed's other mandate, unemployment, which has very slowly fallen since October 2009; that part of the mandate is even fuzzier than the inflation part, but a falling unemployment rate in single digits and low inflation around 2% of the CPI is what most Fed officials want, I assume.

Note that my main point doesn't require that inflation is EVER controllable by the central bank. All I'm saying is that there's no stylised fact that central banks in the developing world are generally trying to create inflation and are failing. It is logically possible that the inflation is the result of non-monetary policy factors, but the facts that would suggest a Keynesian liquidity trap (sub-target inflation in spite of monetary policy action) just aren't there.

That doesn't mean that there is no shortfall of AD, because I don't think that these central banks respective nominal mandates are very good. The problem is not that central banks in developing are failing to hit their mandates, because almost none of them are failing, but rather that these mandates have problems.

Brad DeLong says I "speak sooth" (which I learned, by Googling, means he thinks I'm basically right). So one person gets what I'm saying, anyway.

But maybe I need to do another post, in part responding to Scott Sumner, and thinking about David Beckworth's post.

Nick, I did my own analysis of this question here. I think this would be contractionary regardless of the magnitude of the increase in p.
See here: http://bit.ly/RVnba0

K,

"If the US is still paying 1985 vintage coupons on par-callable bonds then the US government is an idiot."

They are an idiot because - those interest payments might cause them to go bankrupt? Maybe you haven't been paying attention. See Abba Lerner:

http://en.wikipedia.org/wiki/Functional_finance

"Principles of 'sound finance' apply to individuals. They make sense for households and businesses, but do not apply to the governments of sovereign states, capable of issuing money."

Increasing NGDP would likely encourage the government to borrow and spend more. This is because with higher NGDP, there are more tax revenues, and with more tax revenues, there is more opportunity to lend to the government to earn interest.

Raising NGDP isn't the solution. The solution is much more painful.

Frank,

The yield on 30-year US government bonds during the years just preceding 1985 was around 12%. If there were any still outstanding they'd mature within 3 years. So the US government *could* pay 36% percent of notional in interest over the next three years *or* they could just call the bonds at par, and issue new 3-year bonds at 0.3% interest or so. If they were incapable of making the right choice that would be gross negligence as well as idiocy.

Nick,

You may "speak sooth" but you haven't addressed the fact that the Fed controls the rate on t-bills. If you write another post it would be great to address the mechanism of a vigilante attack in a closed economy where government is financed in t-bills. (You, yourself pointed out both that your argument applies in a closed economy and term debt is irrelevant).

Frank,

"The yield on 30-year US government bonds during the years just preceding 1985 was around 12%. If there were any still outstanding they'd mature within 3 years. So the US government *could* pay 36% percent of notional in interest over the next three years *or* they could just call the bonds at par, and issue new 3-year bonds at 0.3% interest or so. If they were incapable of making the right choice that would be gross negligence as well as idiocy."

Or they could cut other spending and pay the notional 36% percent interest or they could raise tax revenue to cover the notional 36% interest. You are confusing private finance that has interest rate sensitivity with public finance which does not. Again there is this gross misconception that economics is about right versus wrong or good guys versus bad guys.

K: I was (implicitly) assuming the debt was all very short term Tbills all along. If it were all 30 year bonds, I would have needed to talk about the lag between the Fed increasing the rate of interest and debt service costs increasing. In other words, I really don't understand your point.

Frank,

Yes, they can *choose* to just give 36% extra to the bond holders, rather than spending it on valuable projects or cutting taxes. But I'm guessing they don't.

Nick,

My point is what I said yesterday at 12:25PM:

Why does higher government debt cause a vigilante attack? Lets say the policy rate is at equilibrium and then there is a surprise increase in government spending/debt, but the Fed keeps the policy rate, and therefore the t-bill rate, fixed. Also assume the money was wasted. 

What is it that suddenly makes bond holders feel the real rate is too low and they now want to invest/consume more? That would make sense if you can establish an automatic link between more debt and higher expected inflation. But what if more debt causes people to expect disinflation? This is also a possible (dis)equilibrium. Then the demand for savings would increase and the policy rate would have to decline (see 1930's or Japan). 

What is the necessary link from higher government debt to a higher natural rate?

*Or*...

What if the governments inflation resolve is unquestioned? No more or less inflation expected as a result of the additional debt. The only impact of the deficit spending is that future taxes will now have to be higher leading to reduced investment incentive and raised current spending incentive. Why doesn't the natural rate decline?

Good post Nick. One of these days you're going to realize you need to stay stuff like, "Printing up more $100 bills is good *within limits*," or "Investors attacking our bonds is good *within limits*." But I'll take what I can get.

You might want to skim my reaction to Krugman. We're basically saying the same thing, but I think mine has more drama near the end.

Oops I meant "One of these days you're going to realize you DON'T need to say..." I.e. printing money out of thin air doesn't make us richer, period, and a bond attack is bad, period.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad