[Second UPDATE: Gene Callahan wants to be clear that he did acknowledge that he had initially misunderstood Nick Rowe’s model. Now that he gets what Rowe was describing, Gene concedes that it is possible, but points out (correctly) that we can achieve the same impact on everyone in the model through government tax-and-transfer actions, rather than using government bonds at all.]
[NOTE: I have made a few edits for typos and one misstatement–I wrote “9 generations” instead of “9 time periods” when I had Nick spell out his wager–since I first posted this. The below is just the post-correction version.]
Wow this whole debt debate has truly been a formative moment in my life. Seriously, I’m not kidding. I thought that with the crystal clear precision of a specific numerical example, all of our bickering would fall away. Don’t get me wrong: I wasn’t expecting everyone to fold up shop and say, “You win this time Murphy, but I’ll be back!” However, I expected people to say, “Whoa…You’re right Bob, this is a lot more subtle than I had initially been thinking about it. But now, using the very helpful diagrammatic framework you’ve developed, let me show you the points I have been trying to make…” (BTW, some of you have done that with me over email, and I appreciate it. I do it all for the fans.)
Fair warning: This is going to be a long post. If you are a casual reader, you may want to skip it. But I am writing this one primarily for PhD-trained economists, because even you guys (with a few possible exceptions) aren’t seeing all the moving parts on this stuff.
Now let me be clear upfront: As of this moment I am backtracking a bit, and now think that Steve Landsburg’s specific points have all been correct, except when he says that Krugman has been shedding light rather than confusion on this topic. As I will soon explain, Krugman, Dean Baker, Abba Lerner, et al. were thinking about this the wrong way–and so was I in my two book treatments of it. I think Steve has been mystified by my commentary thus far, because Steve has been thinking about these things mathematically, as opposed to using economic intuition. So Since Steve has a formal proof in his head (I’m guessing), he knows that what he’s been saying all along has been correct, and he is just trying to figure out how presumably smart guys like Nick Rowe and I can be making such boneheaded blunders.
Thus, my task in the present post is to wind back the clock, and try to remind people of what they would have thought last week before I entered the fray, and amplified the viewpoints of Don Boudreaux and Nick Rowe. To put it simply, I don’t believe any of you when you looked at my numerical example and thought it was no big whoop. No, it is a huge whoop.
Let me close this preamble with two final observations to try to convince my PhD-holding readers why you should really give me a chance to make my case:
(1) I haven’t been just dabbling on this issue while I go take a smoke break or something. (I don’t smoke, kids, and neither should you. You don’t need nicotine to be cool.) No, Nick Rowe’s original post caused me physical discomfort. I literally couldn’t get to sleep the first night I read it, because I was so sure he was wrong, and yet, there was some little voice warning me that I needed to actually write it out and prove that he was wrong. And when I started trying to do that, it hit me like a brick wall that he was right, and I had been thinking about this stuff the wrong way for at least 10 years. I have probably devoted a good 20 hours to this issue since then.
At my current level of understanding, I realize that there are at least three things tripping people up: (a) Assuming the generations don’t overlap. (b) Assuming that bonds passed on to the next generation must be bequeathed, rather than sold. (c) Assuming government tax-and-spend decisions over time have to be subgame perfect, as opposed to merely constituting a Nash equilibrium. If you have no idea what I’m talking about on each of these points, then I am pretty confident that you don’t realize how subtle this stuff is.
(2) I’m not picking on Gene Callahan, but I just want to explain why I think it’s worth all of our time to catch our breath, step back, and give Nick and me a fairer shake. Nick originally wrote his post, the one that knocked me on my keister. Gene’s initial reaction was to say that the model was “logically incoherent.” Then, I came up with an Excel spreadsheet to clarify exactly what Nick was saying in his largely verbal post. Go read his original post, in light of my spreadsheet, and you will see that I am truly just spelling out what he was describing there.
So: Gene originally said Nick’s post was logically incoherent, I spell it out with a clear numerical example, and then Gene…acts like my example just proves the points he’s been making all along. Nope, sorry, that’s a personal blog foul. It shows that this issue is extremely subtle, and people aren’t giving it the attention that I am saying it deserves. (To be clear, it’s possible Gene et al. have been making other, valid points that Nick and I are missing. But it’s hard to keep track of everything, when people keep saying stuff about our position that is simply not true. Let’s get this stuff down first, and then we can move on.)
Suppose one month ago, Nick Rowe put a riddle/challenge on his blog that went a little something like this:
I have developed a simple model of an economy spanning 9 time periods that has the following properties:
==> People live for two periods.
==> There are two people alive each period.
==> There are 9 total time periods.
==> It is a pure endowment economy with one good (“apples”) with no physical carrying of apples through time.
==> People have standard preferences: They just care about how many apples they eat, they want to smooth consumption over time, etc. There is no altruism and no envy. “No funny stuff.”
==> There is perfect certainty and everyone’s actions form a best-response to everybody else’s actions, including the government’s policies. I.e. it is a Nash equilibrium.
==> If people so choose, they can lend apples to the government, to be repaid the next period (with the revenues coming from either taxation or from borrowing again). However, these bond deals with the government are purely voluntary; a given individual will only do it, if s/he prefers the revised consumption flow to the original endowment flow.
==> There is a single act of taxation in period 9. At that time, the government will take 100 apples from someone, and give them right back to the same person.
==> Relative to the alternate timeline in which everyone consumes his/her endowment each time period, in this instance a person who is alive in time period 1 achieves higher utility while a person who is alive in time period 9 achieves lower utility. Everyone else in the model is either the same or better off, compared to the endowment baseline.
So my wager: I claim that I really am thinking of a model that satisfies all of the above criteria. If you doubt me, then you put up $500 to my $5,000. Call my bluff if you dare. After I get all takers, I will reveal whether I’m bluffing or if I actually have an example satisfying all of the above conditions. I promise, there’s no funny stuff like a person who is worse off because he finds government deficits psychologically distasteful or something goofy like that. Nope, I am either bluffing or I will present a numerical example that you will agree, fits the bill.
Now then, who would have taken Nick up on this offer? I think Gene Callahan would have, because he would have concluded that Nick is being logically incoherent; no such model could possibly exist.
Abba Lerner would have, assuming Nick Rowe is correctly characterizing Lerner’s position.
Paul Krugman would have, because Krugman thinks about this stuff by saying: “Second — and this is the point almost nobody seems to get — an over-borrowed family owes money to someone else; U.S. debt is, to a large extent, money we owe to ourselves.”
Matt Yglesias would have, because he wrote:
A few recent Krugman posts making the point that domestically held government debt doesn’t create a “burden” on the country in the sense of draining it of real resources have proven surprisingly controversial. I think the point can maybe be more clearly made in reverse. Imagine a country with a balanced budget and a large outstanding debt, all of which is held domestically. Tax revenue, in other words, exceeds spending on programs but the extra revenue is needed to pay down the existing debt. If the stock of debt is burdening the country, then it ought to be able to enrich itself by defaulting. Will that work?
Well, no. Certainly a default could set the stage for enriching specific people, since it would create budget room for a tax cut or new spending on a shiny supertrain. But the funds flowing into the pockets of taxpayers or train-builders would be coming out of the pockets of bondholders. A government borrowing money from its own citizens doesn’t gain access to any resources that wouldn’t have been available by conscripting them or raising taxes, and by the same token a country doesn’t enrich itself by refusing to make promised interest payments to its own citizens. It’s only when borrowing from or repaying foreigners that the country as a whole is gaining or losing access to real resources.
And for for sure Dean Baker would have fallen into Rowe’s trap, because Baker wrote:
As a country we cannot impose huge debt burdens on our children. It is impossible, at least if we are referring to government debt. The reason is simple: at one point we will all be dead. That means that the ownership of our debt will be passed on to our children. If we have some huge thousand trillion dollar debt that is owed to our children, then how have we imposed a burden on them? There is a distributional issue — Bill Gates’ children may own all the debt — but that is within generations, not between generations. As a group, our children’s well-being will be determined by the productivity of the economy (which Brooks complained about earlier), the state of the physical and social infrastructure and the environment.
One can make the point that much of the debt is owned by foreigners, but this is a result of our trade deficit, which is in turn caused by the over-valued dollar.
I would have fallen for Rowe’s trap too. Let me walk us all through why I would have, so we can see how our intuition–displayed clearly by Krugman and Baker above, despite Landsburg’s attempt to rehabilitate them–is wrong.
First, I would have dealt with the weird thing about the government in period 9 taking 100 apples from somebody, then handing them right back. I would think, “That’s clearly a wash. It doesn’t hurt or help a person to take apples away and give them right back. If the government took apples from the person and gave them to the other person alive in period 9, then there would be a redistribution, but Nick said that’s not what happens.”
Just to check myself, I might say, “Hey Nick! Can you give me a little more info? How does the other person in period 9 feel about the two scenarios?”
Nick would say, “Sure I’m a sport, Bob. I’ll tell you that–assuming I’m not just bluffing here–I am picturing a model where the other person in period 9 is indifferent between the two scenarios.”
Now I would be sure that I was getting $5,000 from Nick. I would reason this way: “Clearly there is nothing going on in period 9. The government doesn’t change either person’s consumption from the endowment alternative, so it is equivalent to a world where there was no taxing at all. So I’ll just move that issue to the side, and pretend there’s no taxation. Now, Nick said that was the only instance of taxation, and that any other activities are voluntary, and people will only undertake if it makes them better off. Since I’ve just proven that the government isn’t hurting anybody in period 9, and that is the only possible door through which the government in this model could hurt anybody, then clearly Nick is bluffing.”
But wait, $500 is a lot of money after all. So I would think about it some more, and come up with a completely independent argument to convince myself that Nick was bluffing. I would reason this way: “Nick claims that somebody in period 1 is better off compared to the endowment outcome, while somebody in period 9 is worse off, and the only taxing activity occurs in period 9. But he says this is a pure endowment economy, people only live two time periods, and there’s no altruism nor envy. So even if I didn’t already have my first argument, I’ve got this consideration to give me confidence: It is physically impossible for anything done in period 1, to have any connection to something in period 9. There are several generations of people who live and die in between the guy in period 1 who allegedly benefits, and the guy in period 9 who allegedly is the only one who suffers. There isn’t a time machine, for crying out loud, to suck apples out of period 9 and into the mouth of a person in period 1. So for sure, Nick has to be bluffing. I have not one, but two knockdown, bulletproof arguments.”
So I would have taken Nick’s wager, and he would have said, “Thanks Bob. BOOM!”:
And, after recovering from the shock/horror/joy, I would have gotten out my checkbook. And, given the quotes I presented above–and their similarity to my “knockdown, bulletproof arguments”–Nick would have collected from Gene Callahan, Abba Lerner, Paul Krugman, Dean Baker, and Matt Yglesias. He also would have collected from at least 20 people who have chimed in on all of our blog posts on this topic.
Now I’m not sure if Landsburg would have lost the bet. But I will say this: If I had read Landsburg’s blog post beforehand, it would have encouraged me to take Nick’s wager. Specifically, here are two of the points Landsburg presented in his summary post:
9. As far as future generations are impoverished by the greed of the current generation, their main beef should be not with deficit financing but with the underlying greed. Indeed, a greedy current generation is perfectly capable of impoverishing future generations without deficit spending, by depleting their inheritances.
13. On the other hand, even with credit constraints, deficit finance does not allow an entire generation to increase the burden on its descendants beyond what it could have done anyway (and in that sense is nothing like a time machine). It only allows some families to increase the burden on their descendants. The greatest damage one generation can inflict on the next is to consume everything in sight, and this is always possible without deficit finance.
So like I say, reading Landsburg on this issue would have encouraged me to fall into the clutches of the clever Canuck. Points (9) and (13) would have made me think, “Just so, Steve, good job clarifying all this stuff. As Nick makes clear in his alleged example, there is no altruism–so inheritances are always going to be zero, anyway–and it is a physical necessity that each time period, all the apples are consumed by somebody. So I am now even more sure, that Nick is full of it.”
(I’m not going to make this post any longer than it already is, but let me be clear that I actually think Steve’s points  and  are technically true–even in this specific case–but are highly misleading, as I hope you’ll all agree if you’ve stayed with me this long…)
Yes, it is a huge deal. It’s not merely that Krugman et al. were wrong in letter, they were also wrong in spirit. Look again at that example above: Do you not see how that is EXACTLY what the layperson has in mind, when he says, “It’s not right, us paying for government services that benefit us, by running up big deficits. If we want goodies from the government today, we should have the courage to let them raise taxes and pay for it ourselves, instead of passing the buck down to our kids and grandkids. At some point, that game has to come crashing down, and those poor saps in 100 years are going to be furious at us.”
Despite all of my “bulletproof” arguments, there is like a “wave” that flows down through the generations, from the initial “unfunded” transfer payment in period 1 to Old Al. It’s true, there’s not a time machine, but c’mon, that chart I came up with illustrates perfectly what the average Joe has in mind. In contrast, Krugman et al. have been saying stuff that is not only wrong on the specifics, but is generally misleading and would make the layperson draw the wrong conclusions. Perhaps all of their points work, if we don’t assume overlapping generations, but guess what? That means their result rests crucially on a simplifying assumption that is false, meaning they shouldn’t be trumpeting the result.
Last point to show why this is very important: A lot of you keep saying, “Deficits are functionally equivalent to taxation.” Yes, you’re right, in the sense that I could come up with a complicated scheme with no deficits (involving redistributive taxation in periods 1 through 8 ) to mimic my second scenario above.
But in the real world, it is not at all true that “deficits are equivalent to tax-financed schemes.” Yes, if the government pays old people $1 trillion today, it is “paid for” by redistributing $1 trillion away from young people today (rather than in 100 years). But it certainly makes a big difference to those young people today if they get government bonds in exchange for their present sacrifice of $1 trillion, or if they are simply hit with a tax bill for $1 trillion. The only people who have emphasized this critical point (to my knowledge) are Boudreaux and Rowe, before I came on the scene to be the Thomas Huxley to their Darwin.