Bob Murphy plays with the debt burden

I don't normally do blog posts with just a link to someone else's blog post. We aren't that sort of blog. But I'm going to make an exception in this case. Bob Murphy's latest post on the burden of the debt is very clear, very comprehensive on both sides of the debate, and very funny. It's a fitting capstone to this debate.

That's probably it for me on this subject. But you never know for sure........

OK, just one last slogan: macroeconomics is about PEOPLE, not GDP!

Comments

[Frances Woolley]

Nick: "OK, just one last slogan: macroeconomics is about PEOPLE, not GDP!"

Spoken like a microeconomist!

[JKH]

Haven’t absorbed all those Murphy posts, but it seems to me he’s complicating the core issue by getting into all that interest rate/tax offset detail, and utility measurement. These are interesting complications - but not essential to the main point of difference with Krugman.

The core domestic issue is whether or not a generation sells or bequeaths bonds to the next generation, and whether the final generation that gets taxed has bought the bonds to end up net down, or has hedged its position via inheritance. But with an upfront transfer, somebody in the future bears the net burden, unless the first generation receiving the transfer is also the one that gets taxed.

And the core international issue is that where foreigners hold the debt, the final tax is a net domestic burden for the final generation, while the bonds are irrelevant – similar net burden result, but via a different gross configuration.

[David Andolfatto]

Nick, love the slogan. Of course, I can't help pointing out that that people cannot be located in IS-LM analysis (and the like). They are, however, easily located and emphasized in modern macro theory.

[Nick Rowe]

Frances and David: Yep! Macro can be like micro in that regard. Sometimes the difference between people and GDP doesn't matter. But in this case it does.

David: did you see this old post of mine?

http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/12/the-trade-cycle-vs-is.html

It's a sketch of a business cycle model where GDP is constant, but welfare fluctuates with the cycle. Simply because there's less trade.

JKH: yes and no. I see your point, and it's mostly right. But what Bob did was create a model where he could talk about people's utility. And that matters, because we are not indifferent to when we consume our lifetime consumption. We would rather not eat all our apples when we are young, or when we are old.

[JKH]

Nick,

Yes. I’m no expert on utility, and it looks like he did a very good job of extending your model in that direction. But to the degree that part of this was motivated by a response to Krugman, it appears to me that the utility dimension could be a bit of a stretch as a point of dispute, while the core aspect of intergenerational nominal distribution is a more obvious difference with a blanket “we owe it to ourselves”. That’s all.

[anon]

Why did Murphy only tax one person per generation, until the end? (And even then he inexplicably taxes one person 9 apples and the other person 1.)

[Nick Rowe]

anon: I think it was just to get whole numbers.

[Lorenzo from Oz]

Bob Murphy's post is great. Except that we are not paid in apples, we are paid in money to buy apples and the central bank can drive down money incomes so it becomes much harder to service the debt. The real issue with public debt is the "room to manouevre" issue--(a) how much of current taxes are consumed to service it; (b) how much extra-debt raising ability is available for emergencies; (c) implication of what the debt is spent on for future income (and so future [a] and [b]).

[Determinant]

That's it Nick. The post where you felt bad for posting lots of abstruse theory about money was pushing it, but this is truly the last straw.

I have forwarded this post to the International Macroeconomist Cartel. They will be dispatching an elite team of Free Market NINJA's (No Income, No job, No Assets, they do whatever they are contracted to do) to your office. There they will seize your PhD and well, you too. You will be transported to the Wayward Macroeconomist Re-education Camp for an intensive series of seminars, lectures and tutorials to bring you back to the True Faith (TM). Don't worry, this won't hurt a bit. No, that's not true, it may hurt a bit if you don't tip the NINJAs, good entrepreneurs that they are.

Have a nice vacation. ;)

[Andy Harless]

"That's probably it for me on this subject."

Wait a minute, Nick, you can't quit before you've dealt with the real issue. All this stuff about apples is very interesting, but it misses the critical point: Krugman believes that deficit spending will increase the number of apples produced. Now I understand why it still might result in a burden on future generations, but the argument is different: it's a behavioral argument rather than an accounting one.

[Nick Rowe]

Andy: sure. Deficit financed spending (relative to tax finance of the same spending) will benefit the current generation (absent Ricardian equivalence). Both the direct effect, plus any multiplier effect, in a Keynesian model with deficient-demand unemployment. I don't have anything much to add to that (beyond what we already worked out in comments in the first post).

The benefits could well outweigh the costs. And yep, it's a behavioural question.

My brain isn't coming up with anything especially new on that subject.

[Bob Murphy]

Thanks for the link, Nick, and I too love the slogan.

Andy Harless wrote:

All this stuff about apples is very interesting, but it misses the critical point: Krugman believes that deficit spending will increase the number of apples produced.

I don't think that's "the critical point." I really think Krugman was saying that, in a closed economy with lump sum taxes, government deficits per se can't make future people poorer on net, they can only make some people in the future poorer while making other people in the future richer. This seems obvious when you realize that real output isn't affected by government debt per se.

However, Nick's cute apple example (and now his slogan) shows the fallacy in that line of thinking. So I really do think that that is what Dean Baker, Krugman, and Yglesias were stressing. They weren't saying, "Yep, the debts we run up today will make all of our grandkids poorer, except that the employment generated will offset that tendency and on net will make them richer." No, I think they were saying there's not any impoverishing effect that needs to be counterbalanced at all.

[Nick Rowe]

Bob: I hope you realise how totally honoured you are, to have a whole post on WCI that does nothing except link to your post (and add a slogan)?

(Actually, linking to your post just saved me the hassle of having to go over all that stuff myself, plus, you did a much better, and funnier, job of it than I could have done.)

Yep. There is no way I can read Paul Krugman or Dean Baker as saying anything other than 'the kids are alright; don't worry'.

What really worries me is that monetary policy won't be loosened enough to stop the debt burden slowly climbing for a long time. Not just in the US. And the argument for a big short 'pump-priming' deficit to break out of the recession to stop the eventual debt burden getting bigger is not a stupid argument (though monetary policy would be better).

Google tells me that Andy has been making twitters inviting/provoking a response to this post from PK and Matt Yglesias! Still radio silence. Oh well.

[Greg Ransom]

If your economics is about people, and your economics isn't about appealing to magic with GDP smoke & mirror constructs, then the zkeynesian fantasy of government directed increased value production across time isn't much of a plausibility -- for a host of reasons (we can see when the econ is anout people not GDP) government has little likelihood of producing greater output than it undermines.

It's a question of what people can possibly do -- and how the government could possibly do better, or why we should have any expectation how it possibly would or could.

nick wrote:

"Andy: sure. Deficit financed spending (relative to tax finance of the same spending) will benefit the current generation (absent Ricardian equivalence). Both the direct effect, plus any multiplier effect, in a Keynesian model with deficient-demand unemployment. I don't have anything much to add to that (beyond what we already worked out in comments in the first post).

The benefits could well outweigh the costs. And yep, it's a behavioural question."

[Max]

"(b) how much extra-debt raising ability is available for emergencies"

You have to be careful to define the emergency. An increase in the demand for money is an emergency, but not one that prevents the government from spending money. (Painfully obvious when stated this way, and yet smart economics say dumb things like "the 2008 recession shows that fiscal space is valuable").

[JKH]

On reflection, I think Krugman will shoot all this down by saying none of it changes (pre-tax) income in each period, that he covered the generational redistribution issue in his general distribution comment, and that neither he nor his readership is concerned about utility functions in the context of the more straightforward point he’s making.

Nick, your model works because you’ve defined it according to future generations of people rather than future populations existing in given future time periods, which PK will shoot down as not what his perspective on “future burden” was.

Murphy defined his model according to time periods, but it only works (in his final argument) because of utility functions, which PK will shoot down as an academic refinement and distraction from the simpler point of fact he was making.

But PK is still wrong on the foreign debt holdings issue, the one which Murphy never addressed. And I’m not sure you’ve presented a full argument on it yet.

(Slight accounting note: the redistribution doesn’t change pre-tax income, but it does change discretionary income, with the offset being the repayment of debt principal (which is not an accounting income item). That decomposition actually doesn’t appear in Murphy’s table, since the repayment includes a blend of interest and principal; interest is an income item.)

[JKH]

Sorry, I guess in that case, PK would be right on the foreign debt holding issue as well, because there he's essentially taking a population view (domestic versus foreign) within a time period view, so he's being consistent. There's a net domestic loss due to taxes.

[Jim Rootham]

My key observation in all this is that empirical evidence trumps theory.

[Nick Rowe]

Greg: OK, there are all the normal arguments against aggregation. But in this case, it's not so much a question of whether we aggregate, but how we aggregate. Is the unit of analysis the time period, or is it the cohort?

Max: I would put your point this way: if there's a recession, and you want to use a fiscal policy to fix it, why not use monetary policy instead? If it's an increase in the demand for money, we could print money, buy back bonds, get out of the recession, and actually reduce the future tax burden at the same time.

JKH: I think that's correct (assuming RE is false and no bequests).

Jim: In this particular case, the claim being made, that I am arguing against, is not an empirical claim. It's not even a theoretical claim, in the normal sense. It's the claim that it is logically impossible (under certain assumptions) to impose a burden on future generations. I am saying that is is logically possible. So I only need one hypothetical counterexample to show I'm right.

Whether or not there is in fact a burden on future generations is indeed an empirical matter, and will depend on a lot of things.

[123 (TMDB)]

Most of the people here have become New Austrians without realizing it. Old Austrians argued that the expansion of debt coupled with the behavioural mistakes by businessmen will cause wrong investments to be made. New Austrians argue that the expansion of debt coupled with the behavioural mistakes by households will cause wrong inter-generational resource allocations to be made. To make their argument more relevant to today's political situation, New Austrians ignore behavioural mistakes made when creating private debt, and focus only on mistakes made with the public debt, even though there is no difference - both public and private debt can be used to burden the future generations. Krugman has a model where the debt cycle drives AD when interest rates are close to zero. He wants to restore the growth of debt to trend. Now he has got a new set of critics - the New Austrians.

Behavioural finance should be kept out of the discussion about AD. They belong to the personal finance papers (in the case of New Austrians), and they belong to Harvard Business Review (in the case of Old Austrians). If normal AD is restored, the number of behavioural mistakes will increase. If normal AD is restored, more houses will be built in wrong counties, and more apples will be eaten by the wrong generations. Get over it. Krugman should keep the radio silence.

[Nick Rowe]

TMDB: "Most of the people here have become New Austrians without realizing it."

By "here" do you mean the US, or this blog, or? Interesting take, in either case.

[Greg Ransom]

This is the macro/Keynesian fallacy -- that "output" has value / significance unrelated to relational context -- unrelated to changing production good structures and human valuers.. It's the Ricardian fallacy constantly corrected by Menger, Walras, Jevon's, Wieser, Mises, Hayek, Bohm-Bawerk, Lachmann, Buchanan, and others.

"Value" isn't a substance that you can "output" in a univocal mass "K" or "GDP" or I + G + C or any universal, cardinal and summable "value" extrusion.

"Murphy defined his model according to time periods, but it only works (in his final argument) because of utility functions, which PK will shoot down as an academic refinement and distraction from the simpler point of fact he was making."

[Greg Ransom]

Why is Plato's account of meaning wrong? Because there are no collective "meaning" entities which sum over an aggregate of particular uses that give meaning to our actual uses or a world. The meaning is embedded in the actual uses.

Why is the macroeconomists account of "value output" through time wrong? Because there are no collective "value entities which sum across time over an aggregate of particular valuations and uses, and that assign and endow value to particular uses of goods. The government cannot extrude "value" like donuts out of a machine by summing over the value calculated from the past of what has been dumped ion the machine.

The valuational significance of particular goods is embedded in the particular and changing alternative uses and valuations perceived by particular evaluators in particular places and in particular times.

This isn't about "aggregation" it's about what gives things significance as "output", " consumption goods", "production goods" and what gives any particular item it's valuations status,

The socialist calculation debate comes into DIRECT play. A central planner can't simply deem things to have a fixed and univocal value. And neither can a macroeconomist.

[Max]

"If it's an increase in the demand for money, we could print money, buy back bonds, get out of the recession, and actually reduce the future tax burden at the same time."

Buying (long term) treasury bonds is paradoxical. It is alleged to work by lowering interest rates, but if the market actually thought it was effective, interest rates would go up. If confusion were an economic stimulus, it might work.

Buying risky assets like low grade corporate bonds and stocks has more potential. At worst, the central bank makes a ton of money. But maybe that is the problem - the CB's clients don't like being bypassed?

[Steve Roth]

Okay, so this was all worked out decades ago. Either it is a burden on future generations, or it isn't.

There's the same number of apples in generation L, but less utility because government distorted revealed preferences.

That's the takeaway, right? And the second clause in that sentence is the one we're still debating.

[123 (TMDB)]

Nick,

by "here" I meant "people persuaded by the macroeconomic significance of Nick's overlapping generations criticism". But other meanings of "here" work pretty well too.

So you need to prove that the other ways of boosting AD will avoid the overlapping generations problem, or you are arguing that the stability of AD needs to be balanced against the overlapping generations problem. I am arguing that the other ways of boosting AD have the same problem, as monetary stimulus will induce the creation of more private debt, and behavioural problems will also arise as result of wrong estimates about expected stockmarket returns after the monetary policy boosts the equity values. And it is also wrong to refuse to stabilize AD if you fear the intergenerational problem, as other policy tools are available to achieve the intergenerational justice.

And maybe the real current problem is that the lifetime consumption of current generations is too low. Pre-crisis popularity of mortgage refinancings supports this idea. So the government debt expansion might be helping us to reduce the burden on the current generations.

By the way, I have noticed that you have argued to some of the commenters that the public-private debt distinction is somehow important for this debate. Do you still believe it?

[Nick Rowe]

Steve: no. There's always the same number of apples eaten per period, by assumption. But generation C (in my example) eats fewer apples in its lifetime. Any loss of utility from eating them too soon or too late in its lifetime is over and above that.

[123 (TMDB)]

Nick:"What really worries me is that monetary policy won't be loosened enough to stop the debt burden slowly climbing for a long time. Not just in the US."

Why didn't you worry about the slowly growing burden of private sector debt when the monetary policy was normal? (I know you did, there is something about this in the archives).

Krugman was a market monetarist in 1998. He argued against the fiscal stimulus in Japan. Maybe he has realized that monetary stimulus increases debt too. The main difference is the type of debt created (public vs. private). Which type of debt is more costly? Who were less wrong, Italians (they created a burden on future generations via accumulation of public debt) or Spaniards (achieved the same via housing bubble debt)? I am not sure. As Williamson said, a lot of depends on relative efficiency of collection of private debts vs. collection of taxes. And as the public debt requires a smaller administrative apparatus, maybe the productivity would grow if half of the private credit-card debt could be replaced by the government debt.

[Nick Rowe]

123: "By the way, I have noticed that you have argued to some of the commenters that the public-private debt distinction is somehow important for this debate. Do you still believe it?"

Yes. An individual who takes on debt expects to gain by doing so. And it's his own name on the IOU. When the government borrows, the name on the IOU is left blank, until the government increases taxes. The eventual name on the IOU could be someone yet unborn, and will generally be different from the person who benefits.

Just because government debt has a cost on future generations doesn't necessarily mean we shouldn't use it. It has benefits too, for both current and sometimes future generations. But monetary policy would be better, because it avoids the cost on future generations.

[123 (TMDB)]

Nick: "An individual who takes on debt expects to gain by doing so."
A government that takes on debt expect to gain by doing so too. So the only difference here is the existence of the public choice problem?

"And it's his own name on the IOU."
Not exactly. It is often the name of unborn children, as you can roll over the debt. So the difference is that you don't create a public choice problem for the future generations.

So I still see no difference between the public and private debt except the existence of the public choice problem. Importantly, both types of debt are used to shift spending between generations.

[Nick Rowe]

123: well, it's the people alive today who vote in the government, but some of the people paying the debt aren't born yet.

I can't make my children pay my personal debts.

[Determinant]

That's exactly what happens when your estate is liquidated, Nick. Your heirs inherit both your liabilities AND your assets. The fact that life insurance frequently mitigates this problem does not mean the problem does not exist.

[Nick Rowe]

Determinant: they inherit my net assets, right? And I can't leave negative net assets (I think)? A national debt lets us leave negative net assets to our kids.

(And, under Ricardian equivalence, where people do leave positive bequests to their kids, and anticipate and offset the future burden of the national debt, we have already agreed there is no burden from the national debt.)

[Min]

Determinant: "That's exactly what happens when your estate is liquidated, Nick. Your heirs inherit both your liabilities AND your assets. The fact that life insurance frequently mitigates this problem does not mean the problem does not exist."

Nick Rowe: "They inherit my net assets, right? And I can't leave negative net assets (I think)? A national debt lets us leave negative net assets to our kids."

Gov'ts roll over their debts until the **kids** (both debtors and creditors) die or are conquered. The conquerors, if any, take over the assets and maybe (some of) the debts.

[Determinant]

Sure you can leave debts. You can leave a house and a mortgage to your kids and in Ontario, in order to inherit the house you also need to take over the mortgage. It is entirely possible to inherit a house with negative equity.

[Lorenzo from Oz]

Greg Ranson: This is the macro/Keynesian fallacy -- that "output" has value / significance unrelated to relational context Which is exactly my objection to the concept of 'malinvestment'. What is a good investment is entirely dependent on circumstances. What is a great investment in New York may be a very bad on in Port-au-Prince. What is profitable in high level of economic activity may well not be in low ones.

It is also my objection to blanket condemnation of government output because it is government. Sure, there are good reasons (and empirical evidence) that government production tends to be problematic. But it is also true that its quality varies enormously -- between activities, over times, between polities. Yes, there are problems with GDP valuing government production at cost, but it is not guaranteed to be zero value either. Indeed, there are no guarantees about its value at all. (Which is another argument against fiscal stimulus.)

Max: yes, though Nick's reply is much better.

On public debt: is not the key issue government as intermediary? Yes, debt is both a liability and an asset. An asset to bondholders and a liability to taxpayers (these being overlapping groups). Using public debt-to-GDP ratios as the most common measuring metric is a bit odd; but not completely so, since the thing being borrowed against is the taxing capacity of the government.

Debt is basically a structure of promises. The larger the (public debt promises) structure compared to GDP, the greater the difficulty in keeping the promises (to pay); the more risk that promises will not be kept; and the more difficulty in making further promises (to pay). Hence the interest rate on government bonds vary between polities and over time.

The problem with the welfare state is if it systematically generates more debt (i.e. promises to pay) while also putting pressure on revenue (ability to pay). There is a difference between public debt at 15% of GDP and public debt at 150% of GDP. There is also a difference between debt when income is rising and debt when income is falling (and the more so the more it falls).

Any story about debt that does not include risk rather misses the point, surely. After all, risk is the most single important variable in valuing debt. Because debt is a structure of promises (to pay) and, in valuing promises, what is going to be more important than their expected reliability?

[Shangwen]

Nick, I usually stay far away from these impressive macro debates because I am totally ignorant of macro. But because of that, I'm going to ask you a simple man-in-the-street question, understanding that I may be missing a huge point and just be way off base.

Today we all sit in the midst of many public goods and services for which debt was long ago incurred--hospitals, schools, public-sector pensions, war, bailouts, etc. Is it relevant at all to this discussion what the quality of those goods were? Or is blowing money on war or the GM bailout indifferent, in a macro sense, to simply incurring the debt and not buying anything with it?

I try to limit my comments here mostly to health care. Consider Saskatchewan under Blakeney and Devine in the 80s-90s: they used what was ultimately public debt to buy more inpatient beds per capita than any other province (ultimately many were closed by Romanow). That was wasted money. We went to war in Afghanistan for (in my view) nothing. Our biggest cluster of program expenditures, health services, is grossly inefficient.

If all that debt had been spent with breath-taking wisdom and ingenuity, so that we all lived to a healthy 120, worked until 90, or all our kids math scores were past Singapore's, would that make a difference? I am really just asking a naive question here--maybe the wisdom of the debt-related spending is not the issue here. Or is this just an apples discussion where we don't care how sour or sweet they were?

[Bob Murphy]

Nick Rowe wrote:

Google tells me that Andy has been making twitters inviting/provoking a response to this post from PK and Matt Yglesias! Still radio silence. Oh well.

I would love to hear PK chime in, because he is one of the few people who will recognize in a few seconds why I did the apple model the way I did it. It cuts right to the essence of the dispute. Yes, the other stuff (distortionary taxes, Ricardian Equivalence, an output gap in the present, etc. etc.) are all relevant too, but Krugman et al. had assumed it away. They were saying if you strip away all those things, then "duh!" handing our grandkids some pieces of paper, telling the government at that time to take money from some grandkids and give it to some other grandkids, can't make the grandkids poorer on net.

Even after reading Nick's posts on this, I still couldn't really "see it" until I did that Excel table. And then I stared at the thing in shock. I still catch myself discussing its features carefully, because it shattered 10 years of my intuition on this issue.

So, my point is that I wasn't expecting PK to chime in right away. I think if he ever delved into our debate, he would have been as shocked as I was, since I could tell the arguments he was making were *exactly* the way I was thinking about the issue, before Nick chimed in with his 3-period apple example.

Last thing: Somebody said PK would just say it isn't about utility. I doubt he would do that. If he does, I think most professional economists would know, "OK Krugman lost." Do you really think he's going to try to tell his readers that yes, deficit finance will foist a life on our grandkids that they would prefer not to have, but that we're still not hurting them by doing so? What would that even mean?

[Nick Rowe]

All: I gotta go to bed. And I have a very long very busy day tomorrow. Will try to get back to this, but much later. Sorry.

[Nick Rowe]

Determinant: but if I leave my kids a house with negative equity, they can walk away and refuse to accept the bequest.

Lorenzo: yep. The simple stories we are talking about here do leave a lot of such things out.

Shangwen: you are right. It matters what the government buys when it borrows. If it buys the right assets, there will be an offsetting benefit to future generations, and that could exceed the cost to future generations I'm talking about here.

Rob: I agree. It's not an easy point, but PK will get it quicker than you or I did. He really does need to say at least "OK, there's another way of looking at this...". He can't ignore this and complain about other people's mistakes and ideology.

[J.V. Dubois]

Nick: "They inherit my net assets, right? And I can't leave negative net assets (I think)? A national debt lets us leave negative net assets to our kids."

I think it is even worse. You can force that debt on other children, not your own. Assume your apple economy again, but this time fix the per capita production (apples are produced by labor). Assume that children can not only buy but that they can also inherit assets from their parents. If you now borrow apples from someone else and then die, he will not leave less assets to his children (we assumed that assets can be inherited). By assuming that children indeed inherit at least part of the assets, your decision to borrow and hen die left some future children with less consumption then they would otherwise had. In your model when you ruled out even partial Ricardian equivalence this cannot happen as your decision to consume more would just mean that someone else withing the same generation would consume less. But with mechanism of transferring even virtual assets such as claims on future production to the children for "free" (inheritance).

But anyways, I do not feel any more Austrian now. As Nick said, it is about cost-benefits analysis of debt.

[JKH]

“I think if he (PK) ever delved into our debate, he would have been as shocked as I was”

That’s seems inconceivable to me. He would have to be oblivious, from a mental accounting perspective, never to have thought of the net cash flow difference resulting from the case of a taxed individual (or generation) inheriting bonds versus paying for them. If the issue is assumed to be neutral in the standard economic text interpretation of recent years, then one of those scenarios MUST be non-neutral by inference, IF the assumption is that you are looking at discrete future generations as opposed to generations defined as the consolidation of such discrete subsets at a future point or period of time.

And something that I can’t understand is how any thoughtful economist who has ever interpreted Ricardian equivalence as inheritance in this sense (and it appears they have), would not recognize the distinction between such net cash flow scenarios immediately.

So I think PK is smarter than what you assume in your expectation of his reaction here. I’ve never seen an OLG model in my life that I can recall, but I could figure out the critical difference between inheritance and purchase cash flow effects shortly after seeing Nick’s model. Krugman and Nick are of the same generation, so they may have the same sense of history as to how these things have been taught.

The key to all this is whether you define future generations as discrete generations at a future point in time or over a future period of time, versus defining them as the totality of all such future discrete generations that in fact co-exist at the same future point in time or over the same future period of time.

If you go with the former definition in the context of getting to either the Rowe or Murphy models, then the nature of the problem becomes clear when you consider that an intergenerational bond transaction (sale or inheritance) requires the effective co-existence of two different generations at that future point in time, at least in the sense of the two counterparties to the transaction (even though one is dead in the case of inheritance). That has two consequences. First, it means there is definitely a net cash flow difference to the taxed generation when comparing purchase versus inheritance. Second, it also means that under the alternative interpretation of co-existing future generations, ALL such transactions (purchase or inheritance) net out under accounting consolidation to no net effect. And it’s that latter interpretation that Krugman will say he intended.

So he’ll say he wasn’t distinguishing future generations in isolation, but future generations as they co-exist in the future. And he’ll cover that as intended in the broad sense of how he qualified everything according to the distribution issue. That seals the deal for him right there.

“Somebody said PK would just say it isn't about utility. I doubt he would do that.”

That’s not exactly what I said. I said it would be qualified according to his readership. And inherent in that qualification is the fact that NYT limits his space per blog post. He’s made that point often as it affects his presentation of potentially complex topics. And this latter point is a VERY important fact in both his presentation and how one should interpret it.

PK obviously knows the utility story. But he’ll say several things in response. First, his intended co-existing generational interpretation makes that point moot. Second, while acknowledging the potential for the alternative discrete generation interpretation and the point on utility which pertains to that case, he’ll that the NYT rules on his blog space limitation precluded going into either of those beyond the blanket qualification he made about distribution generally.

And he’s right on the foreign holdings in either interpretation - discrete or co-existing generations.

My own prediction is that he WILL respond, with something along those lines.

[Bob Murphy]

Bob Murphy wrote: “I think if he (PK) ever delved into our debate, he would have been as shocked as I was”

JKH wrote: "That’s seems inconceivable to me.

I don't think that word means what you think it means.

JKH wrote: [Krugman] would have to be oblivious, from a mental accounting perspective, never to have thought of the net cash flow difference resulting from the case of a taxed individual (or generation) inheriting bonds versus paying for them...And something that I can’t understand is how any thoughtful economist who has ever interpreted Ricardian equivalence as inheritance in this sense (and it appears they have), would not recognize the distinction between such net cash flow scenarios immediately.

I understand that once you spot what's going on, it is completely utterly obvious. But nonetheless, it seemed SO COMPELLING to reason like this: "If we abstract away from the possibility of a government deficit changing physical investment patterns, and we assume a closed economy, and we assume lump-sum taxes, then whatever we do in our generation can't possibly affect real GDP in the year 2100. So the people who are living in 2100 collectively cannot be made richer nor poorer by what the national debt is at that time, all that can happen is individuals can be made richer or poorer but there will necessarily be offsetting losses/gains to others alive at that time."

You're right, if you try to reconcile that "insight" with the financial flows, it blows up in your face. But I claim that Krugman never *tried* to so reconcile them, because the "insight" seemed so bulletproof that he never carried it further.

I really do hope he responds, and we'll see if he goes your right. My guess is that if he wants to punt he'll do it like this:

HYPOTHETICAL KRUGMAN REPLY: OK so it seems a lot of people out there tied themselves up in knots thinking about these simple issues of debt. Look, everything I said is true; even my worst critics concede that (after being dragged kicking and screaming). Now it's true, there is a distinction to be made between what our grandkids earn in a given year, versus what our grandkids earn over the course of their lifetimes, and there is an interesting subtlety there that trained economists should be aware of. But in the real world, we have a massive output gap, and the economy is screaming for more demand. Right now government deficits are the only thing that can supply it. There is really nothing there in what these critics say.

[Bob Murphy]

Nick: I just have one nagging doubt, when we say we are confident Krugman was making this elementary mistake. Why did he (and Dean Baker) keep drawing a distinction between government and private debt? If they were using the logic we think they were using, it should be just as applicable to private debt (so long as it's held domestically).

[Rob]

In Nick's initial coverage of this issue Cohort C consumed less apples in its 2-period lifetime than Cohort-A and B. This was interesting but of course the # of apples consumed in each period remained the same. Bob's twist was to add-in a utility-function that allowed him to build a model where no only did a later cohort consume less apples but whole generations would have lower life-time utilty.

My concern with Bob's addition was that the results were very much dependent on the distribution of the endowments and the utility function chosen. Ironically for a Austrian's Bobs model required equality of consumption to demonstrate its result .

I fear PK could challenge the results on the grounds that the model does not reflect the real world and if it was amended to make it do so (less equal distribution of endowment), then the results could no longer be derived.

Even worse: the model can clearly be modified to "prove" govt intervention (both taxes and borrowing) will maximize people's lifetime utility.

[Nick Rowe]

Bob: because I can't force my kids privately to accept a negative bequest from me. They can just say "no". If I am leaving them my house as a bequest, then it's true that the bigger my mortgage the smaller my net bequest. But presumably I know about this effect on my kids' wealth and lifetime consumption, and factor it in to my decision-making like Barro-Ricardo.

I can leave my kids a negative bequest if we all do it collectively, through the national debt. And I will do that if I don't care about my kids. And I might do that if I do care about my kids but am confused into thinking that the national debt is not a burden on my kids.

[123 (TMDB)]

Nick: "because I can't force my kids privately to accept a negative bequest from me."

It is the same with the private and the public debt. Private debt has to be supported by the present value of future profits and rents on the asset side, otherwise future generations will not accept it. Public debt has to be supported by the present value of future taxes on the asset side, otherwise future generations will not accept it.

By the way, did you see my January 10 05:10 PM comment in this thread? I am still wondering why Paul Krugman is no longer a market monetarist. Maybe the similarity between the private and public debt is the reason.

[Min]

Nick Rowe: "I can leave my kids a negative bequest if we all do it collectively, through the national debt."

Could you spell out what you mean by such a negative bequest? I know it is a metaphor, but I don't see how it relates to your previous example of a ballooning debt that becomes unpayable. That's more like a will that gives $1000 to heir B, but directs B to give $2000 to heir A, it seems to me. Thanks. :)

[Yancey Ward]

JKH,

I think it clear that Krugman didn't understand the distinction you outlined here:

IF the assumption is that you are looking at discrete future generations as opposed to generations defined as the consolidation of such discrete subsets at a future point or period of time.

Krugman went out of his way to foreclose one objection before it got raised- he made the point that what he was writing concerned internally held debt so that one could not object to his argument by pointing that future Americans could be made worse off relative to future Chinese. Surely he meant to do that because Americans couldn't, by law, readjust the income flows through taxes to make themselves neutral with regards to the Chinese (in other words, he was trying to foreclose a case where they would not owe it to themselves). If this was an important point to make, and we can assume he thought it so because Krugman did make it, then one has to assume he thought future Americans could readjust income flows from internally held debt to make sure none of them bore a burden because of debt we initiated in the present-i.e., this is just a problem of distribution of income at that time. If Krugman really was just making the assumption in the second part of your quote, that he was considering just the consolidation of all discrete subgroups, then there was never any logical reason to mention the distinction between internally and externally held debt since some Americans in our future but in that future's past benefitted from borrowing from the Chinese, while Americans in that future's present (if the debt is retired through taxes, for example) pay for exactly that benefit, or, in other words, it nets out neutral regardless of whether it was externally/internally held. I assert he was trying to have it both ways, but just didn't understand that was what he was doing.

[Rob]

Either I'm missing something or Bob's model does not show that debt can reduce future utility. It just shows that transfers of any kind reduce present utility for one individual and by definition the average of all present and future utility for everyone.

The only way that the model shows that debt "causes" a future loss of utility is in the sense that a debt transfer creates an obligation for a future transfer. However it is clear that the govt could avoid this further loss of utility by simply defaulting or taxing the bond holders an amount equal to the bond payment.

Further: The only reason a transfer causes a loss of utility to an individual and a fall in the average of present and future utility for everyone is due to way the model is setup - the utility function is maximized when consumption is spread equally between the 2 periods of each persons life , and the endowment is spread equally between both individuals alive at any time. If ether of these assumptions is changed then transfers and borrowing increase utility.

So even if one accepts that the model proves that a debt transfer "causes" a future loss of utility that the govt can't avoid by either defaulting or taxing the bond holders, then one also needs to accept its assumptions (equal income for all individuals , and each individual maximizes utility by consuming equally in all period of his life) for this model to prove anything about the real world.

[Rob]

To generalize from my previous comment:

Govt debt can cause loss of utility in the future only if the transfers it effects deviate from those that would have been needed to maximize utility.

In Bob's model no transfers are needed to increase utility so any that do take place cause a loss of utility.

In an alternative model where transfers between individuals or between generation would improve total utility then government borrowing will only be harmful if it causes deviations from this optimal set of transfers that would have been required.

In a model where savings and investment is allowed then government actions can screw things up both by leading to lower than optimal output as well as lower than optimal distribution of goods. But in the unlikely event that the govt can do the right thing at the right time the outcome will be optimal.

[Bob Murphy]

Rob,

I'm not sure what you are talking about, regarding my model. For example you wrote:

The only way that the model shows that debt "causes" a future loss of utility is in the sense that a debt transfer creates an obligation for a future transfer. However it is clear that the govt could avoid this further loss of utility by simply defaulting or taxing the bond holders an amount equal to the bond payment.

No, debt "causes" a loss of utility because, once set in motion, it necessitates that either (a) the government will default, thus reducing the utility (relative to no debt scenario) of whatever sucker rolled over the debt or (b) it taxes future people to service the debt. The only way debt would not "cause" a loss of utility in the future is if everybody could keep rolling it over and nobody were ever taxed, not even to cover interest payments.

The stuff you are saying about the sensitivity of the result to my assumptions is wrong. Just think about it: People are worse off if they get taxed for something that they aren't reaping any benefits from. It's a pretty simple result. Krugman et al. got lost in thinking that in the future, people taxed to service the debt will be helping the bondholders who are getting paid, but that's not a wash, since the bondholders earlier in their lives reduced consumption in order to get into the position of holding that debt. (Or, they might have, unless Ricardian Equivalence is perfectly true.)

[Bob Murphy]

Nick Rowe: Maybe you're right, but the thing with the govt/private debt distinction is that it sounded like Dean Baker was leaving open the possibility that *private* debt could hurt future generations. Remember, his point was that debt couldn't hurt our grandkids. But he said "at least government debt."

So the thing you're focusing on, is actually correct, right? *Private* debt can't hurt future generations, because they can refuse to accept it unless there are offsetting assets to accompany it?

[JKH]

Bob Murphy,

"I don't think that word means what you think it means."

Interesting point.

Perhaps its conceivable as a hypothetical, but inconceivable as an actual.

[JKH]

Bob Murphy,

Your hypothetical Krugman reply does sound a LOT like him.

Will be interesting to see something.

[Rob]

On the first point : Do you agree that any transfer of any kind in your model will reduce average utils per person compared to what it would be with no transfers ? So a bond issued in period 1 will reduce average utility on everyone. Move to period 2. If the govt wants to maximize future average utility from period 2 on it has to default on the debt, or pay the debt and tax the receiver an equal amount. failure to do so will leave the bond-holder (old Bob) better off but reduce the average utils for everyone else in the model who lives in period 2 or later. The same thing will happen in every period - the more transfers the govt does the worse the average utils become because no transfers is always the best option. I'm not explaining this very well but I don't think it matters because I think this is just a symptom of the much more general issue I raised in my second post (and I welcome your comments on that)

On your second point : I agree in the real world tax is bad. But in your model it is not bad(if I understand the utility function) if it equalizes an individuals consumption over 2 periods. For example if the distribution over everyone's lifetime was - period 1: 150 period 2 : 50, if the govt taxed them in period 1 and gave them a transfer in period 2 t the combined tax and transfer would increased their lifetime utils.

I apologize in advance in the (statistically probable) case that I am wrong here and wasting your time.

[Reverend Moon]

@JKH
It's from the Princess Bride
[Vizzini has just cut the rope The Dread Pirate Roberts is climbing up]
Vizzini: HE DIDN'T FALL? INCONCEIVABLE.
Inigo Montoya: You keep using that word. I do not think it means what you think it means.
Share this quote

[Nick Rowe]

Min: "Could you spell out what you mean by such a negative bequest?"

Suppose I die with zero assets, but an unpaid debt. Can my creditor come after my kids to make them repay the debt?

I'm not at all sure of my legal history, but I think that in the olden days in England they could. My kids were liable for my unpaid debts after my death.

Nowadays they can't. If I die with unpaid debts (and no assets), too bad for my creditors. My kids can walk away from my old debts So people will not lend me money if they don't expect me to repay before I die (or leave sufficient assets when I die to cover those debts).

So, as a private person, I can't just borrow what I want, spend it all, and say my kids will cover it. But I can vote for a government that does that for me.

[Nick Rowe]

Bob: "So the thing you're focusing on, is actually correct, right? *Private* debt can't hurt future generations, because they can refuse to accept it unless there are offsetting assets to accompany it?"

Yes.

Unless there are some sort of externalities so that my borrowing causes a financial crisis and a recession that leaves my kids unemployed. Maybe that's what Dean baker meant?

[JKH]

Yancey Ward,

I think I understand the point you’re making, but would disagree. I suppose I’m softening up again on the possibility that Krugman just didn’t think through the numbers, but I don’t think this is a persuasive argument for that.

Suppose Krugman in good faith was arguing the case of consolidated discrete subgroups at some future point in time or some future period of time. So as an example of that, suppose you have the final cohort buying bonds from the previous cohort and subsequently being taxed to pay the bonds down. On a discrete basis, the final cohort ends up short cash as a result of those three transactions. But on a consolidated basis, since the previous cohort is a counterparty to one of the transactions, the two cohorts together ended up flat cash over the period that covered these three transactions. That difference is due to the effect of taking the consolidated view, which has been my assumption for the Krugman interpretation.

As you say, Krugman “foreclosed” on the case of bonds held by foreigners.

But in fact the foreign case corresponding to this example results in a net short cash position for the taxpaying generation under either the consolidated or discrete interpretation. (I do assume that the definition of the relevant generation in either case includes only domestic taxpayers – not foreign taxpayers, and that “we” does not include foreigners, whatever the precise context for “we” turns out to be.) The result is that choice between the “we owe it to ourselves” consolidated interpretation and the Rowe/Murphy discrete generation interpretation doesn’t matter to the correct interpretation of the foreign bond case. The foreign bond case becomes a net burden for the domestic taxpayer in either case.

Assume the consolidated interpretation in the foreign held bond case. So the final foreign cohort buys the bonds from a prior foreign cohort, and the government taxes the domestic sector. The result is that the domestic sector is net short cash, since the bond related cash flow nets out entirely within the foreign sector.

Assume the discrete generation interpretation. This changes the interpretation within the foreign sector, at the option of the modeller I suppose, but that has no bearing on the burden faced by the domestic sector, which is still net short cash. So the result is the same as the consolidated interpretation.

Which is to say that the foreign sector scenario is irrelevant to the issue of assessing the likely applicability of consolidated versus discrete in the case of Krugman.

That’s the way I see it, but it’s also quite possible I really haven’t understood your point. What I’m saying is that the nature of the foreign bond case should have no bearing on whether one chooses consolidated rather than discrete as a ruse to get to the “we owe it to ourselves” conclusion.

[Min]

Nick Rowe: "So, as a private person, I can't just borrow what I want, spend it all, and say my kids will cover it. But I can vote for a government that does that for me."

Who are the creditors in your gov't analogy?

As I said, I don't see how this gov't analogy fits your earlier scenario.

[Bob Murphy]

Rob,

I was never making an argument about looking at "average utils per person." I don't believe in interpersonal utility comparisons. I used the same utility function for everyone to make it simple, but I could just as easily multiplied each successive person's function by a bigger and bigger number and gotten the same result. I was never comparing a given person's utils with somebody else's utils; rather I was looking at how many utils a person got under the deficit scenario, versus the endowment scenario.

Also, you're right that if the endowment had a person start out with (150, 50), then a tax and transfer system that transformed it into (100, 100) would make the person better off. However, the only reason the government would be necessary, is if people in the private market couldn't lend to each other for one period. So if someone tried to justify taxing for that reason, we could say, "Of course, we have to assume private bonds don't exist."

(Now Samuelson et al. *do* make an argument about the benefits of a sustainable Ponzi scheme a la Social Security, which is kind of like what you might be saying.)

[Nick Rowe]

Min: "Who are the creditors in your gov't analogy?"

The kids themselves (or someone else's kids). They buy the bonds off me and the other old guys, but my kids owe it to your kids, and your kids owe it to mine. So they paid for something that they then owe to themselves.

Bob: "Also, you're right that if the endowment had a person start out with (150, 50), then a tax and transfer system that transformed it into (100, 100) would make the person better off. However, the only reason the government would be necessary, is if people in the private market couldn't lend to each other for one period. So if someone tried to justify taxing for that reason, we could say, "Of course, we have to assume private bonds don't exist.""

In a 2 period OLG model (young/old) private bonds cannot solve this problem. The young people want to lend, but who do they lend to? The old can't repay them next period, because they will be dead. This is where samuelson comes in. It could be government doing a sustainable Ponzi debt (sustainable because the equilibrium interest rate would be negative). Or it could be that individuals pick up shiny stones and sell them when they are old. If shiny stones are the only durable good, they will be very valuable as a savings vehicle.

[Rob]

I shouldn't have used the phrase "average utility" but I think the same point can be made in a different (and simpler) way.

You would agree that in a free market with no tax or government bonds there would never be any transfers in your model (people would be happy to lend at a big enough interest rate, but noone would borrow at those rates). For this reason if the govt carries out any transfers it will make at least one person worse off and if the transfer is via debt that person will by necessity will be in the future. I take this to be your point (though you make a point in your model of showing that many generations in a row can have sub-optimal utility , but that always depends upon the govt making additional transfers to make that happen.)

If we have an unequal distribution of endowment then transfers are possible that make some people better off and none worse off as you have acknowledged. These transfers can be via bond or via tax. There are also possible transfers in this model that will make at least one person worse off (for example: if the govt sells the youngers a bond for transfer to the then old, only to then tax them to pay off the bond in the next generation). The point being that each model has an optimum set of transfers that will improve both parties utility. If the govt carries out these transfers (via bonds or taxes) noone loses. If it carries out different transfers then at least one persons utility will drop. In your model the set of transfers that will improve at least one persons utility and decreases no-ones is empty.

"the only reason the government would be necessary, is if people in the private market couldn't lend to each other". This is sort of my point. People will always try and maximize the present value of their current and future income steams. In a free market they will work out the optimal transfers so that everyone (at least ex ante) wins. The reason that govt borrowing is bad is that the govt will screw the transfers up and end up making at least one person in the future worse off (except in your model where all transfers do that). If by a miracle the govt could simulate the same set of transfers via borrowing and tax that the private sector would achieve via borrowing and trade then govt borrowing would not be a burden on the future.

[Rob]

To restate:

If transfers can improve Pareto efficiency then in principal there is no reason why the govt should not be able to facilitate those transfers via bond-financed debt - but in reality one may doubt the likelihood of the govt achieving this better than a free-market.

[Bob Murphy]

Rob I'm not trying to be difficult but at this point I don't know what you are getting at. If we have a model where one guy grows apples and another guy grows oranges, and we say for some reason they can't trade voluntarily, then a government tax scheme could make them both happier.

[Min]

Moi: "Who are the creditors in your gov't analogy?"

Nick Rowe: "The kids themselves (or someone else's kids). They buy the bonds off me and the other old guys, but my kids owe it to your kids, and your kids owe it to mine. So they paid for something that they then owe to themselves."

So the creditors are heirs and the debtors are heirs. Which is different from the private debt, where the creditors are not heirs and the heirs are possibly debtors.

BTW, around 20 years ago I heard on public radio about people inheriting debt in India, which had led to debt peonage because of ancestral debts. I don't know if the law has changed since then.

[Nick Rowe]

Bob: I think (not 100% sure) you are not getting my &Rob's point.

Let me try it this way, because it doesn't necessarily involve government.

Suppose there are exactly 1,000 people in each cohort. Suppose one of them comes up with an idea: he makes exactly 1,000 special engraved disks, that cannot be copied. He calls a meeting, gives everyone in his cohort one disk and says: "when you are old, sell your disk to one person in the next cohort for 50 apples, and tell him to do the same when he is old". The disk is like a chain letter, except you only pass on one copy, not several, and everyone only gets one copy.

As long as nobody breaks the chain, everyone in all cohorts is better off. The first cohort is much better off, because it eats (150,100) instead of (150,50). But all subsequent cohorts are better off too, because they eat (100,100) instead of (150,50).

Min: interesting about India. That's probably illegal now (unless the government does it).

[Rob]

Bob,

I am saying that in that scenario the trade and the govt transfer would achieve the same result and you would have to bring other factors in to show why the trade is better.

It would be easy to construct a model where govt actions exactly simulates the transfers involved in a debt transaction that would benefit both parries.

Instead of:
A lends to B, B consumes more (he has a family to feed), and in the next period he pays B who needs to pay for his retirement (obviously have dropped the 2 period OLG constraint here).
we have:
Govt issues a bond and sells it to A, and the govt makes a transfer to B. In a later period the govt taxes B and uses the money to repay A who can now retire.

Now I can think of many reasons why the free-market version is better than the govt version (mainly around revealed preference). But within the constraints of the discussion we have been having doesn't the second option show that govt debt sometimes does not cause loss of utility?

[Andy Harless]

"There is no way I can read Paul Krugman or Dean Baker as saying anything other than 'the kids are alright; don't worry'.

I can read Krugman and Baker as saying, "The kids are alright; don't worry, because the increase in production as a result of deficit spending has enabled us to increase their bequests by as much as we have increased the debt, and what matters is not the debt itself but how much is produced and consumed." (This isn't Ricardian Equivalence; it's a matter of people leaving larger bequests because they have more income, not because they expect their descendants to have less.)

One could imagine asking them the question, "Did the Reagan deficits (i.e., when we weren't in a liquidity trap) make our children poorer (ignoring the part borrowed from abroad)?" and the answer could be, "No, the deficits didn't make our children poorer, but the additional consumption that resulted made our children poorer."

It would be nice to hear what they actually have to say.

[Peter]

I have to say I really appreciated Bob Murphy's post because it highlights a point that I've struggled to stress with non-economists, namely that we do not disagree on the methodology, models, or outcomes of well defined models. Where we do disagree is on two key points 1) the assumptions, and 2) the focal point of the outcome. (And the focal point of the outcome is usually tied to subjective preferences or underlying assumptions.)
The assumptions are necessary for the solving of the model, but they often get burried and lost deep in the bowels of the model construction and occassionaly forgotten.

[Lorenzo from Oz]

So, what I take from this discussion is that public debt that is being serviced is not a burden on future generations.

(Of course, in policy land, what really matters is whether a given level of debt that can be serviced: but that is not the issue being discussed here.)

[Nick Rowe]

Lorenzo: Not really. It is the future servicing of the debt, in paying taxes to cover interest and/or pay down the debt, that will be a burden on future generations.