Suppose you wanted to take milk away from people on the east coast, and give it to people on the west coast. But you don't have any way of transporting the milk quickly enough to stop it spoiling before it gets from one coast to the other.
Here's how you do it:
You take milk away from people on the east coast. You move it a few kilometres west, then swap it for fresh milk from the people living a few kilometres inland from the east coast. Then move that second batch of milk a few kilometres west, and repeat. And keep on repeating until you get to the west coast.
Suppose you wanted to take milk away from people who won't be born until 100 years in the future, and give it to people alive today. But you don't have a time-machine.
Here's how you do it:
You borrow milk from people who are young and give it to the people you want to give it to. You wait a few years, then borrow some more milk from people who are young, and use it to repay the milk you borrowed from the people who were young but are now old. And keep on repeating until you get to the people who were born 100 years after you started. You don't borrow milk from them; you just take it.
Now read what Paul Krugman said:
"First, however, let me suggest that the phrasing in terms of “future
generations” can easily become a trap. It’s quite possible that debt can
raise the consumption of one generation and reduce the consumption of
the next generation during the period when members of both generations are still alive." (italics in original).
Worthwhile Canadian Initiative 
How is private debt different? It has pretty much the same distributional properties (age-wide) as public debt. Why such selective asymmetry in the topic of burden?
Sergei:
With private debt, you cannot take the milk from the people 100 years from now.
With public debt, you can (and do.)
Bill, why? Do you say that once in a while banking system assets go down to zero and after that it all restarts again? Or that we have generation specific banks, i.e. each generation borrows/lends exclusively to itself?
in PK words, you could take from A and give to B, take from B and give to C, etc., all the way to Zed.
Sergei: I can’t force my kids to inherit my debts when I die. I cannot leave them a negative bequest. They can just walk away. (The law was once different, IIRC, and if I could force my kids to inherit my debts, it would be just the same.)
@Sergei:
You can reduce the amount of goods you left to your heirs by consuming more. This is a way to raise your consumption at the expense of the next generations.
But if you want to do what the state do at the individual level, heirs should be denied the possibility of renouncing their heritage when the defunct has a negative equity. That was the case in many societies (and it is probably still the case in some parts of the world).
Nick, banks do everything to make sure that you pay off . For one thing banks will require that you have life insurance. Finally, equity of banks still belongs to future generations once you “walked off”. So you personally might be able to “walk off” but it does not negate the issue that some future generation will take your bill.
Nick, this is a brilliant example that clearly demonstrates that you are right. I hope you get the public acknowledgment you deserve from the blogosphere – if not, at least private emails acknowledging that you are right.
Once that’s done perhaps attention will shift to questions that more explicitly acknowledge the potential for inter-generational transfers such as: are we in a world where r < g, is r < g sustainable (or can we exploit it with ‘trills’), are we in a world where bonds are mostly bequeathed (I don’t think we should use the term Ricardian equivalence, because it is possible to write a model with bequests and without Ricardian equivalence), how productive is current government spending, and, following up on rsj, what are the inter-generational consequences of monetary policy, etc.
Great post, again!
primed: Thanks! Agreed with the rest of what you say. (I had to edit your comment to put spaces each side of the < and > things. Otherwise it goes all html.)
Nick, this is a brilliant example that clearly demonstrates that you are right. I hope you get the public acknowledgment you deserve from the blogosphere – if not, at least private emails acknowledging that you are right.
Once that’s done perhaps attention will shift to questions that more explicitly acknowledge the potential for inter-generational transfers such as: are we in a world where r < g, is the r < g sustainable (and if so, can we exploit it with something such as ‘trills’), are we in a world with bequests (I don’t like calling it Ricardian Equivalence because it is possible to write a model with bequests and without Ricardian Equivalence), what are the actual benefits of current government spending, and following up on rsj, what are the inter-generational consequences of monetary policy, etc.
Great post again!
(sorry I forgot about html tags in the previous comment – admins, please delete that comment and this stuff in parentheses)
“You don’t borrow milk from them; you just take it.”
Right.
It all depends on the assumption of future taxes.
I assume Krugman, Delong, et al. are resisting with arguments as to why future taxes won’t be necessary?
Or do they not agree with your more basic point that future taxes = future burden?
DeLong:
“Seems to me it would be much more productive right now to worry about how do we maintain normal levels of net investment in a high government debt post-interest rate normalization environment than to propose sending the economy back into recession in order to reduce government debt accumulation”
i.e. he’s resisting the idea of reducing debt accumulation, which means he’s resisting the idea of future taxes – and not necessarily resisting your idea that future taxes = future burden
so he’s resisting your counterexample, but maybe not the logic of your counterexample
It looks like Krugman doesn’t even consider the scenario of future taxes to pay off debt.
His entire interpretation seems to exclude that scenario.
Isn’t that the problem there?
This is all just a repeat from a year ago.
I think their argument is implicitly r < g under current circumstances, not that it always is or will be.
I think Nick’s point all along has been that it is wrong to claim as a general proposition that debt is not a burden on future generations. There are scenarios where it is, and he has shown this.
It looks like DeLong’s response is to agree with that (roughly), but that those scenarios should not be the priority concern in today’s environment. But Nick is not addressing the question of priority concerns; he’s addressing the erroneous claim as a general proposition that debt is not a burden on future generations. There are examples where it is.
Krugman’s response is to omit consideration of those scenarios entirely in defending the general proposition that debt is not a burden on future generations.
To round out, the MMT approach (my impression only) is that from an operational standpoint, there is no necessary burden on the current generation in terms of servicing or rolling over the debt incurred by previous generations. But that’s purely operational – based on the argument that a fiat currency issuing government can’t be “insolvent”. That said, MMT does not preclude payment of taxes under certain circumstances (inflation). And Nick’s examples involving payment of taxes to pay off debt are not really disputable as examples of burden, even though MMT would not recommend that as a strategy under most circumstances.
The way the system works now, I don’t believe that milk is being borrowed. It is something else.
Instead of young and old, what about having wealth/income inequality (the rich who have positive real earnings based on their monthly budget and the lower/middle class who have negative real earnings based on their monthly budget)?
Totally agree, JKH.
I think the MMT approach has always been just to insist that whatever goods and services are produced at a time are consumed by people alive at the time they are produced.
I think Nick’s basic point is that “people alive at time t” designates groups of different generational cohorts, and that final goods and services produced at time t could always be disproportionally consumed by the older cohorts and disproportionally produced by the younger cohorts, which is fine as long as the growth in production over time means generations eventually always receive more from those following them than they gave to those who preceded them. But if that rate of growth declines, generations in the declining phase could lose out. And if there is a final generation, then that generation loses out obviously.
A point I have been trying to make is that the goods and services produced at time t includes not just consumption goods, but capital. And the capital goods are disproportionally produced from the wealth of the older generations and disproportionally benefit the younger generations. Also, if Nick’s mechanism of the young purchasing debt from the old is not operative, then any production financed by borrowing is accomplished by successive redistributions within generations, not by transfers from younger to older.
Sergei said: “How is private debt different? It has pretty much the same distributional properties (age-wide) as public debt. Why such selective asymmetry in the topic of burden?”
Agreed, but I’d rather focus on wealth/income inequality instead of young/old. How is private debt different (and the same) as public debt is a really good question.
Nick’s post said: “Sergei: I can’t force my kids to inherit my debts when I die. I cannot leave them a negative bequest. They can just walk away. (The law was once different, IIRC, and if I could force my kids to inherit my debts, it would be just the same.)”
They walk away and that causes a recession or depression. The gov’t says no “depression”. The gov’t goes into debt to make up for the private debt. Any difference?
Dan Kervick said: “Also, if Nick’s mechanism of the young purchasing debt from the old is not operative, then any production financed by borrowing is accomplished by successive redistributions within generations, not by transfers from younger to older.”
I think so. Need to dig into the details of people’s monthly budgets to see what is really going on.
Nick,
Last year’s discussion is coming back.
But I think I was slightly wrong above about Krugman, as follows. He says:
“OK, you can see what’s coming: a debt inherited from the past is, in effect, simply a rule requiring that one group of people — the people who didn’t inherit bonds from their parents — make a transfer to another group, the people who did. It has distributional effects, but it does not in any direct sense make the country poorer.”
Maybe he does consider taxes there – by analogous construction from his previous example. Maybe he is assuming that those who don’t have the bonds are taxed (the transfer) to pay off those who do hold them. I’m not sure now.
But in interpreting your example, he’s not considering the projection of potential purchasing power currently available to two adjacent future generations considered together. You are saying that there will be a future generation that will incur a burden because it won’t have the option of selling its bonds to the subsequent generation. The bonds are gone. No option. So the potential for that purchasing power is gone. That is a burden relative to the scenario in which they do have that option.
In that regard, Krugman is incomplete in his analysis. He is looking at the tax cash flows but overlooking the importance of the bonds.
The tax transfers money from one bank account to another. And looking at that part of the transaction alone, he’s right. That part is neutral in terms of the wealth of future generations.
But he’s forgetting that the bonds are gone. And from an inter-generational perspective, the bonds can be sold to the next generation in exchange for additional purchasing power. So the future generation that no longer has that option has incurred a burden.
The loss of the option to sell the bonds for additional purchasing power is a permanent loss for the generation that held the bonds at the time of the tax. They no longer have that additional purchasing power.
However, the loss of that option is neutral for the subsequent generation that would have bought them. The subsequent generation would have been swapping current purchasing power for presumed future purchasing power of their own in a similar way in their own future – which is neutral. As it is, they’ll now be doing neither, which is also neutral. So the net result is that the effect would be neutral on that generation either way.
So the full result is that there is a burden for the generation that no longer hold the bonds to sell to the subsequent one. And the burden for all future generations consists only of the burden to that generation that can no longer sell the bonds.
I think this incidence of burden is reflected in some way by the fact that the taxes destroy net financial assets (NFA), to use the MMT term.
In a sense, Krugman has captured the neutrality of the liquidity aspect of the transaction – the fact that the tax cash flow alone just redistributes money between two bank accounts.
But he hasn’t captured the non-neutrality of the capital aspect. The capital aspect is inherent in the private sector value of NFA. Taxes destroy that value. The bonds no longer exist. So they can’t be used to transfer purchasing power to the generation that would have held them. They are poorer in that sense.
Krugman wrote:
“Suppose that after the 2016 election President Santorum tries to buy senior support by giving every American over 65 a gift of newly printed government bonds; then the over-65 generation will be made richer, and everyone under 65 will be made poorer (duh).”
This seems to be Krugman’s logic: if X is given the right to consume more (and exercises it), then Y must consume less, therefore Y is poorer.
But if Y is only consuming less now in exchange for the promise of being able to consume more later (i.e. he bought a bond from X), he’s not automatically poorer. He’s poorer if the government taxes away the bond so he can’t sell it to Z (somebody not born yet).
Duh?
Nick, are you agreeing with Krugman, or arguing that the year 2112 is somehow analogous to the easternmost point of the landmass? If the latter then you may have a point; he seems unwilling to allow that time must have a stop.
It seems the counter to your argument is, yes, you can transfer milk from the east coast to the west and time travel, but it really is only a sequence of transfers between neighbors and generations alive at the same time, so we really don’t gain anything more from extending it through space and time. In fact we lose something, because even if that is our intent, we have no means assuring ourselves it is possible and all our expectations and predictions of the future may fail. So you really are better off looking at each individual transfer when it occurs and whether it makes sense at the time.
1:26 +
In other words, the cash flow effect and bank account transfer that Krugman describes as neutral in connection with a tax to pay off a bond – is no different than the cash flow effect in the case of a bank account transfer due to the purchase of a bond that was still outstanding.
Common sense suggests those are two quite different events, but he really hasn’t made any distinction that applies to the tax situation. So something’s fishy, and that has to do with what I described as the capital value of the bond, above, which is its value in commanding additional purchasing power while it is still “alive” for the generation that holds it and can sell it to the next generation.
I sense that the milk example should somehow self-evidently refute Krugman. But I do not get it.
In Krugman’s reference frame, where he basically considers a living generation against not yet living generations, there is no net debt at any point of the process you describe. The milk that is given to the old is paid for, just not via taxes, but directly: you take it from the young and give it to the old. You just decide to repeat the whole process again and again. Nothing changes for the next iteration of the process: the young and old again will have their basic quantities of milk as it corresponds to the unregulated distribution, and then you redistribute again, but directly pay for it.
And as I see it, that also holds true for the generation in a hundred years: you take it from the young, to give it to the old. There is no net debt, and net consumption is unchanged, too. As far as I can see, this is exactly what Krugman describes.
What do I get wrong?
Martin,
Some day you get to the west coast.
Sorry, scratch that “this is exactly what Krugman describes”, I forgot to delete it.
Nick, why don’t you make it clear that you’re talking about an interest burden, not a debt burden.
Why would anyone issue debt in an economy with no growth?
If economic growth failed to keep up with interest rates what should we do? Shouldn’t we either give the bondholders a haircut or try to inflate away the difference?
You keep assuming we’d illogically guarantee payments in excess of what the economy produce. Why would we do that? This is why your model doesn’t make sense and no one can figure out what you’re talking about.
If we were targeting NGDP and inflation outpaced growth long-term: how is this different? Wouldn’t older generations consume more?
If a bubble causes people to think they’re wealthier than they really are so they consume more than they’re worth: how is this different?
K: “Some day you get to the west coast.”
It’s the east coast which has the problem, in Nick’s somewhat confusing parable. He should, I think, start with the looters on the west coast, with BC taking milk from Alberta, which takes it from Saskatchewan, etc., until at last some poor soul in Labrador is left short. How this relates to borrowing from future gnerations is unclear to me. If the world ends in 2112 then I don’t think it’s debt we should be worrying about.
Kevin: I don’t find Paul Krugman’s post very clear. But I read him as saying that it is not possible to use debt to make transfers between two cohorts unless those two cohorts are alive at the same time. And I am disagreeing with him.
I am pretty much on the same page as JKH@12.23. (JKH and I might disagree about the conditions under which it would and would not be necessary to raise future taxes to service the debt, but that is a secondary issue.)
Hi K,
why do I get to the west coast some day – in time? When is your future time limit? Is there a relation to the Maya calendar? Is that really the argument “There will be a last generation, and that one won’t get their milk back!”? Really?
And again, that’s not even what I am saying – I don’t even manage to see where debt is created in the above example. Not even if you assume overlapping generations.
Say there are two living generations, A and B, who have a total of x liters of milk, and that total is constant for every time period under consideration. So now you redistribute and A gets all, and B has nothing left (as far as I can see, it does not really matter if you claim to take it from generation B or generation B+98 – this is what happens). Net consumption is unchanged, but surely B is worse off than A in that time frame – that’s Krugman’s duh! Now A dies off, and C is born. Again there is a total of x liters of milk… and that’s it.
So, now, as I understand the story line above (without getting why this should be an implication) B thinks there is a “debt”. Personally, I think there is no debt – B has just been forced to pay for the additional consumption of A; that’s unfair, but they did pay, and there is no debt. But anyway, B has the feeling, on moral grounds, that they have been deprived of something that somebody should give them back, now. So, as A is dead, they want it from C. So let that happen: again, the total amount of milk consumed is x. And again, within a time frame where both generations are living, one is better of than the other. Duh! Where does that contradict Krugman? What does this even have to do with Krugman?
What is more is that if you consider the total living time of one single generation, they consume exactly the same, be it with or without redistribution (save generation A, but I didn’t like them in the first place). There is a strong feeling that, as A got away with more consumption than any of the following generations, that some day, a future generation has to consume less and is therefore worse off. But I do not see how the above example shows that, except you assume mankind to go extinct after generation A+99.
Yes, the alphabet does stop, as do the US when you go in one direction for long enough. But time? I do not see that this amounts to a suitable analogy. All the example shows is that you get a generation worse of if you define a time limit, in this case a generation in a hundred years, and/or treat that one generation differently from all the others by singling it out and insisting that it is that one generation you are borrowing from. Why this should be so and why this should illustrate anything pertaining to what Krugman said escapes me.
In your original example when group a receives the apples from the government you have to remember that generation a is buying the bonds, and therefore loses 100 apples in consumption, making net change in apples for that generation 0, similarly, when generation c gets paid by the government for the bonds it receives 121 apples back making net change in consumption 0 as well.
Martin, Kevin:
You don’t necessarily get to the West Coast, but that’s a different issue. Nick acknowledges that government debt can be a stable Ponzi scheme if the growth rate exceeds the interest rate (the “Samuelson” case that he mentions in other posts). However, the argument here is concerned with the case where the interest rate exceeds the growth rate, in which case you do get to the West Coast eventually, in the sense that the debt will eventually have to be paid off or it will keep rising without limit and start to dwarf the real economy.
@ Andy Harless
That may well be, but there is nothing in the above example that shows what you are explaining.
this almost makes less sense than the apple example. Neither seems to demonstrate that the next generation’s income is less – just that there is more distributional inequality in a no-growth model, since interest is flowing to the apple/milk rentiers.
Sergei: “How is private debt different? It has pretty much the same distributional properties (age-wide) as public debt.”
Really? I would have thought that the older generation was overall the creditors of the younger generation, who have debts like mortgages and student loans. No?
Thanks. 🙂
Martin,
Nothing, including humanity, lasts forever. I don’t see the difference whether it’s 100 years or a billion. Either way
1) somebody’s going to get screwed
2) they can’t consent to it
K,
I sersiously think we are talking past each other – and it may well be my fault.
I don’t mean to say that Nick Rowe is wrong on the issue. But when I try to go through the examlpe he presented above, all I see is an illustration of Krugman’s point. There is no inherited debt in the common sense, only a – sort of deliberate – rule to redistribute within future generations. That’s exactly what Krugman says.
Also, I really do not think that an end to mankind is what Rowe alludes to. Perhaps it is something like what Andy Harless mentions. But this point is simply not in the above example.
Maybe I just didn’t get it! In this case I hope I could make clear how I read the example as to provide a basis for correction.
Also, you seem to be concerned with intergenerational equity, and there is an argument to be made in your favor, read for example John Quiggin’s recent paper on “Equity between Overlapping Generations” (mostly concerned with a zero pure rate of social time preference):
http://ideas.repec.org/a/bla/jpbect/v14y2012i2p273-283.html
However, I do not see that this is what Rowe is concerned with – or, for that matter, that this is the topic at hand. I do not say that Rowe is wrong, or something. But also, I do not see how his conclusion is borne out by his model as stated above.
Nick, in your counter-example, you are again confusing the “real” and the “financial”. Of course milk consumed by somebody cannot be consumed by somebody else. And of course the last generation that does not live to see its sweet retirement (because of an apocalypse or whatever) will be “worse off” (yes, their lifes were shorter!). But that is not the point.
The point is that selling an IOU to somebody DOES NOT reduce the supply of finance for anybody else. The supply of money (or debt) is unlimited, therefore, as far as you are far from full employment (of people and real resources), which is, in a free market economy, except for times of global war, the NORM, REAL investment is NEVER conditional on FINANCIAL saving. NEVER EVER.
You seem not to be able to understand this insight of Keynes (and of many others). An accounting identity is not a production possibility frontier. There is something called extensive growth. Without extensive growth, interest does not really make much sense. So your other (apple) counter-example is not very good either. I think commenter Hunter Pritchett rightly notices that you simply summmon the first 100 apples from out of nowhere as an asset without a liability. But that cannot be so. It always takes (at least) two to enter into a debt contract. The debtor and the creditor. By definition, the two sides of the transaction net to zero.
This! And if now Krugman won’t admit that he is wrong, if he will pass this opportunity to show that he is interested in honest debate and something called “truth”, if he refuses to teach those thousands of people following him whom he just misguided, he loses any right to ever utter the word “zombie” or to accuse anyone – Paul Ryan included – from being intellectually dishonest, fraud or whatever. Because he will become one.
Can I get anyone to agree with this claim: Nick has not displayed an example in which the issuance of debt is itself a burden on the future generation. The ultimate cost born by the future generation in his example is not due to the debt issuance, but is due to the fact that, in addition to the debts that are being issued in each generation, the generations are also engaged in a Ponzi process in which each generation successfully induces the younger generation to buy the apple-bonds of the previous generation. And like all non-infinite Ponzi schemes, the folks at the end of the line end up holding the bag. But this has nothing to do with the fact that the things that are being sold in the Ponzi scheme are bonds.
Nick, I have read the thread (following Krugman’s link).
Here is how I understand your thesis. You assert that intergenerational transfers are growing due to the interest on the debt. Yes, it’s true.
But the way to combat it is to tax one generation the sum equivalent to the debt increase and to transfer the taxes to the next generation.
Theoretically, the level of consumption stays constant.
Practically, there will be frictions and discrepancies but we are talking about the concept here.
Regards
Dan, it’s not the debt, but the interest on the debt.
Buying debt is like a loan, if the economy grows then the loan should pay, if it doesn’t it should not.
Nick is giving buyers of debt a guaranteed return on a bad loan.
(He also is portraying debt as unjust transfer payments. I think this is confusing people.)
(I also object to calling a successfully growing economy a “Ponzi” scheme. With this terminology the Volcker recession is “normal,” but the postwar era “weird.” He’s trying to push NGDP targeting, by spreading FUD about fiscal policy, but if inflation outpaces growth under NDGP targeting you still get a “burden.”)
Andy Harless: “the argument here is concerned with the case where the interest rate exceeds the growth rate….”
Maybe that’s the argument we should be having, but I don’t see Nick discussing that.
K: “Nothing, including humanity, lasts forever.”
If Nick has a finite horizon in mind, fine. But in Brad DeLong’s thread he invokes the infinite-horizon OLG model. That’s why his coast-to-coast milk relay seems like a lousy analogy to me.
Dan: that’s like saying; jumping out the window is not what hurts; it’s hitting the ground. (Of course, if r < g, you never do need hit the ground.)
Nick Rowe: “that’s like saying; jumping out the window is not what hurts; it’s hitting the ground.”
As my high school physics teacher liked to say, “It’s not the fall that kills, it’s the sudden stop.” 😉
Boundary conditions matter.
“Suppose you wanted to take milk away from people on the east coast, and give it to people on the west coast. But you don’t have any way of transporting the milk quickly enough to stop it spoiling before it gets from one coast to the other.
“Here’s how you do it:
“You take milk away from people on the east coast. You move it a few kilometres west, then swap it for fresh milk from the people living a few kilometres inland from the east coast. Then move that second batch of milk a few kilometres west, and repeat. And keep on repeating until you get to the west coast.”
The analogy with paying backwards in time is backwards. You should start at the west coast and end up at the east coast. In real life that is difficult. Like it used to be spending a bank note from a Georgia bank in Chicago in the 19th century.
I see taxes discussed here.
Does the “distribution” of taxes matter?