We have been witnessing a somewhat muted round two of the debate over the intergenerational effects of fiscal deficits. Paul Krugman got the ball rolling by enunciating the conventional economic wisdom that “no, debt does not mean that we’re stealing from future generations.” This incensed Nick Rowe, who insisted that, through the vehicle of debt, “we can take apples out of the mouths of yet unborn future generations, and eat them ourselves.” And Roger Farmer chimed in even more strongly: “An increase in government debt always places a burden on future generations.”
I should emphasize that this is not about what use government money is put to; everyone understands that government investment, if it’s productive, can make future people better off, debt or no. And it’s not about Keynesian countercyclical policy, which in a world of hysteresis can also have long-lasting positive effects. No, this debate abstracts from all of this and asks, taken by itself, does a fiscal deficit impose costs on future generations? Krugman says no, that debt is money we owe to ourselves in every generation; the payments and receipts cancel out. Rowe and Farmer say Krugman is confusing time periods with people: the transfers obviously offset at each moment in time, but the payers and recipients can be from different generations.
So let’s sort this out. First, we should frame the question just as Rowe and Farmer do: it’s not about time periods, which is trivial, but generations that overlap within time periods. And let’s also take the issue of what the money is spent on off the table too, since it could be financed either out of taxes or bond issues. Just examine debt’s effects on intergenerational distribution in an economic test tube.
Consider a world that functions like this: (1) All people are economically indistinguishable except for their age. There are no economic classes, nations, or any other categories of significance. (2) People buy bonds for income-smoothing in an uncertain world. The present value of the bond is its price which is equal to the discounted stream of future payments. (3) When they circulate in the secondary market, bonds are always purchased from older people by younger people, never the other way around. (4) People cleverly plan their lifetime earnings and consumption so that, at the moment of their death, the expected value of their bond holdings is zero. Over a large enough population, collectively, it will be zero.
In this world the government throws a big party, financed by a bond issue. The current generation finds itself liable for higher taxes for debt service, but at the same time the possessor of these same bonds. That’s a wash. But they do get the party, so they come out ahead.
What happens next is that some of these oldsters die, and some new babes are born. Collectively the oldsters leave nothing for the newcomers except the tax liability. As more and more of the initial generation get older and see the shadow of the Reaper in the mirror, they sell their bonds to the young. The young acquire the bonds, but they also part with their money; that’s a wash. But they still have the tax liability, so they are worse off than they would be had there been no party and no debt. Moral of the story, the Greatest Borrowing Generation was able to enjoy extra benefits at the expense of Generation Next. And you could demonstrate that the Nexters would be able to borrow from the following generation in the same manner, kicking the distributional can down the river of time, to really mix a metaphor. Thus we would witness “time travel” in the sense proposed by Nick Rowe.
But there’s another world. In this one people are exactly the same as before, differentiated only by age. Bonds are still fairly priced; buying or selling them has no effect on expected wealth. But we drop the rule that says that only young people can buy bonds from old people, and we especially drop the rule that say that, on balance, people plan their lives so that no bonds are bequeathed to the next generation.
The party remains the same and so does the initial bond issue. But when it comes time for Generation Next to take the stage there are important differences. First, the Nexters are receiving many of their bonds for free. This is a wealth transfer from the older generation to the young. Depending on the size of the bequest relative to the number of bonds still in the hands of the aging party veterans it can be partially, fully or overfully offsetting.
But that’s not all. The Greatest Borrowers, even as they approach decrepitude, may purchase additional bonds from the Nexters. That’s a wash directly, but it can eventually result in even larger bequests.
Note that in both worlds it remains the case that at each moment in time assets are identical to liabilities, so no distributional effects across time periods is possible, but generations overlap in time, and it’s possible for resources to be shifted from young to old or old to young. The question is whether this generational transfer is actually taking place and in which direction.
So which world is it? The Rowe-Farmer world is clearly a special case, with no bequests and a strict age structure for bond purchases and sales. (The latter constraint is more important for Rowe than Farmer, but I’ll leave that aside.) It is absolutely true that it is possible to model a special case in which they are right and Krugman is wrong.
But in the general case the direction of the transfer is unknown: it could go from younger to older or older to younger or be too small to notice. And, in case you’ve been reading your Piketty, we do in fact live in a world of bequests. Meanwhile, according to the Fed’s regular survey of household finances, accumulation of financial wealth continues monolithically right up until retirement, so there must be a lot of oldsters buying bonds from youngsters (or disproportionate purchases of new issues). In general, then, Krugman is right.
A lot of economic wisdom boils down to knowing whether you’re dealing with a special case or a general one.