Reply to Wren-Lewis on Inter-Generational Transfer

People, both economists and non-economists alike, seem to have a nasty and surprising habit of forgetting that time travel is impossible, at least when it comes to the economy. Professor Simon Wren-Lewis outlines a simple model in which, by instituting and then ending a pension scheme, goods are redistributed from a generation at time T+10 to a generation at time T, but the problem with this is that it requires time travel to be true.

Think about what “goods are redistributed” means. “Goods” refers to real goods, like hamburgers, cars, computers, gasoline, etc. “Redistributed” means taken from one group and given to another. Suppose the good is a hamburger, and it’s getting redistributed from the generation at T+10 to the one at T today. I don’t know about you, but my first question is, how far in the future is “T+10?” Ten generations would be about 300 years, so the implication here is that the hamburger I eat today got taken away from somebody 300 years in the future. (My next question is, how did it survive the trip??)

Of course that’s not what happens in a pension scheme, and the simple reason is that, to the best of our knowledge so far, time travel is still impossible. If I use my pension check to buy a hamburger today, it might be true that I’m taking a hamburger out of somebody’s mouth, but that somebody is alive today. They must be, or else there would be no possibility of them eating my hamburger instead of me. So how do pension schemes actually work?

In WL’s scheme, in each period both the young and the old produce 50 goods (perhaps 50 hamburgers?) but the young pay 10 goods into the pension scheme and the old withdraw 10 goods from the pension scheme. In this non-monetary economy, it really is ok to think of this as hamburgers: at any one period, the young are producing 50 hamburgers but eating only 40, while the old are producing 50 hamburgers but eating 60. (Not much of a pension scheme if you ask me. Personally, I’d like my retired parents to be producing zero. But maybe that’s old-fashioned…).

So the “redistribution” part should be pretty clear: 10 hamburgers get redistributed from the current young, to the current old. There is no time travel. When time passes and young become the old, then there will still be no time travel: hamburgers will be redistributed from the youth at that time to the elderly at that time.

What about if the pension scheme stops? At T+10, the youth both produce 50 and receive 50, and the elderly both produce and receive 50. There is no redistribution. It’s true that the old are getting the shaft: they produced more than they consumed when young, and were expecting to be rewarded for it when old. It’s also true that the elderly at time T did better than expected, since they consumed more than they produced in their old age, but weren’t being rewarded for anything they did in their youth. But under no circumstances did hamburgers travel 300 years into the past from the screwed generation at T+10 to the lucky ducks at T. In reality, there were a series of redistributions, each one from the current youth to their contemporary elders. The first generation to receive this was able to experience the winning side but not the losing side, the last generation was forced to bear the losing side without enjoying the winning side, while every generation in between got to be on both sides at some point (and if the scheme never ended, then there would never be a generation that would only face the losing side). There is inter-generational transfer here (many times, from youth to elderly), but it is always between generations alive at the same time, and there is absolutely no inter-temporal transfer, because time travel is impossible. What’s more, in no sense is the generation at time T “borrowing” from the one at T+10, or in any way causing their misfortune, for the simple reason that it is the generations alive at T+10 that decide whether to continue the scheme or not, not the by-then-long-dead generation that was alive at T.

However, there are even bigger problems with the model that Wren-Lewis presents next, which stem not only from the time travel mistake, but also from pretending that there’s no money in the economy. His story might be mostly plausibly correct if the government accepted taxes in kind, but it doesn’t, and that matters. Let’s dig in.

The problems begin in the first sentence: “suppose government taxes both the old and young by 10 each period, and transforms this 20 into public goods.” The fact is that you can’t pay your taxes in hamburgers, the IRS will turn you away (and laugh at you). If we re-think this tax into a monetary tax, then we’re stuck with the question of where the government got the 20 hamburgers of public goods it gives out. Alternatively, we could think about what we actually call it when people give hamburgers to the government: unless it’s a donation, usually the government gives money to the hamburger-virtuoso in exchange for the beef patty, and we call that government spending. By trying to box monetary operations into a non-monetary economy, Wren-Lewis has mixed up government spending and taxing.

Let’s quickly use the insights of MMT to re-write this first sentence into what it should really be: first the government imposes a tax of $100 on both young and old, then the young and old sell 10 hamburgers to the government at $10 each, then they use those $100 to pay the $100 tax. Now in possession of 20 hamburgers, the government does…something…to make this a public good somehow. Note that we’ve assumed a balanced budget here, but in fact WL hasn’t made it clear: he did say that the government ends up with 10 goods from each generation, which implies that the government gave out 10 goods worth of money, but he didn’t say anything at all about how much the government received in tax money. If each generation desired to save $40, then the government would be able to obtain $100 worth of goods using only a $60 tax.

Continuing, “in period T the government says that the young need pay no taxes, but will instead give 10 goods in exchange for a paper asset” — but remember, the IRS doesn’t take hamburgers, and we’ve already named the kind of exchange that involves the government giving out paper in exchange for goods: government spending. WL is clearly NOT saying that monetary taxes will be zero, he’s saying that taxes in kind will be zero. Except in real life they’re always zero, because governments don’t do that. In other words, there is absolutely no difference between this second scenario and the previous one, even though Wren-Lewis has convinced himself that there is because he is referring to government spending as “taxes.” If we read WL’s next two words, naming the paper asset as “government debt,” we see that Wren-Lewis has inadvertently channeled a key MMT insight: currency IS government debt. WL states that this debt can be redeemed to obtain private goods from the government in the future; that’s not wrong, but in real life the government doesn’t offer very many private goods for sale (maybe, postage stamps? crap from the Congressional gift shop?). Instead, the primary way those government IOUs are redeemed is when settling debts owed to the government, like tax liabilities.

Ok, to be fair to Wren-Lewis, there is a difference between the first scenario and the second. In the first, we have no way to know what the level of monetary taxes had to be, it depends on the population’s desire to accumulate government IOUs. But in this second case, WL tells us that in period T+1, the (formerly) young will be able to use the government debt to purchase 10 hamburgers from the government. The fact that they still have all $100 in period T+1 implies that the level of monetary taxes was indeed zero. So, our overview of fiscal policy at time T? Government is buying 10 hamburgers each from the young and the old; the young are paying no taxes, while presumably the old are paying some taxes (though we have no way to know how much, it could be anywhere from 0 to 100% of the money that government paid them for the hamburgers, depending on the ageds’ desire to accumulate IOUs). That favors the young, in my book.

Then the pension scheme continues. Because the youth were previously paid for producing 10 hamburgers but not taxed on the income, “in period T+1 this allows them (the now old) to consume 50 private goods rather than 40: the 40 it produces less tax and the 10 it now gets from the government by selling the debt. Their total consumption of private goods has increased from 80 to 90.” Here I think it’s worth pointing out that this is the complete inverse of a normal pension scheme. Typically, we think of a social security or a retirement fund as the government providing money to the elderly so that they can buy goods from the private sector; here, the government has provided money (government debt) to the youth, so that they can save it to buy goods from the government when they’re old. Why Wren-Lewis has inverted the normal scheme, I can’t say. (It might also be worth noting that WL’s scheme is basically voluntary: usually when government spends to buy hamburgers, it will buy from anybody, not just the young. And when it sells things, it will sell to anybody who pays, not just the old. Note too that in the story, the youth save the money until retirement, but there’s nothing forcing them to do that.)

In any case, in order to provide the extra 10 goods to the elderly at T+1, the government must purchase 10 more from the youth, or in Simon’s words, “[government] says to the young: either you pay 20 rather than 10 in taxes, or you can buy this government debt for 10.” Of course, as you and I know, there is no “pay 20 [goods] rather than 10 in taxes,” because taxes in kind do not exist in the real world. The only real choice is government spending. “ They now consume 30 in private goods in T+1,” or in other words, in real terms, the current youth are producing more than they consume, so that the current elderly can consume more than they produce. We notice too that the generation that was old when the pension began did better than every other generation (lifetime consumption of 90 instead of 80), because it enjoyed the winning side in old age but did not bear the losing side in its youth.

WL concludes the logic by examining the ending, noting that the generation just after the program ends gets the shaft (lifetime consumption 70 instead of 80). But he reaches the wrong conclusion by due to both the time travel fallacy and his mistaken definitions of taxes and debt: “we have a clear redistribution of 10 from the young in period T+10 to the young in period T enacted by the government issuing debt in period T.” In truth, we don’t. We have a series of redistributions, in real terms, from the current youth to their contemporary elderly. The elderly in period T+11 do not receive a redistribution from the youth in T+11, but that has nothing to do with choices made in period T. The people from T are long-dead, they didn’t choose to end the pension scheme. The people alive at T+11 did that to themselves.

But even apart from the time travel fallacy, there’s a problem with attributing this series of redistributions to “the government issuing debt.” The fact is that what Simon is defining as “government debt” here is not at all what most people mean by “government debt.” This is caused by the fact that WL is pretending we have a non-monetary economy and that the government taxes in kind. In such a world, “government debt” would unequivocally refer to IOUs that promised goods. But in our monetary world, people take “government debt” to refer to the Treasury Securities that the government sells when it spends more money than it taxes, and these are quite different. What Wren-Lewis has termed “government issuing debt” is what most people would more normally call “government spending,” ie. government giving out a piece of paper in exchange for goods. What most people mean by “government debt” doesn’t even factor into the picture in WL’s story. (I suppose the government in the story could offer to swap some of the outstanding paper debt for some interest-bearing bonds, and then we’d have a “national debt,” but as stated in his article right now there is none.) If Wren-Lewis was trying to tell us that “government debt” causes inter-generational transfer, then he failed because he mis-defined government debt. On the other hand, if he was trying to tell us that government spending can favor one currently-alive generation over another…well, didn’t everybody know that already?

The irony here is, as an MMTer, I agree with Simon’s narrative that government paper in exchange for goods, (what is usually termed “currency,”) really ought to be considered “government debt.” It’s an IOU, which will be redeemed when that paper is used to settle a payment to the government, probably a tax payment (but who knows, maybe a payment for crap from the gift shop). But how many people on the street are counting currency as part of the national debt? Does Wren-Lewis even do that? This kind of government debt is most certainly NOT what politicians and the voters are concerned about when they talk about national debt, nor is this kind of “issuing debt” what they mean when they protest “government borrowing.”

To sum it all up? Professor Wren-Lewis set out to show that government debt can transfer resources from future generations to present ones. But this is incorrect, and in fact all he has shown is that government spending and taxing today can transfer resources between generations already alive today. Future generations will get to make choices about distribution in the future, just like we make them about distribution today. At least until we invent time travel.