Does Government Debt Impose a Burden on Future Generations/Periods/People? #12,143

I think (after a lot of effort) that I’ve internalized Nick Rowe’s modeling of this question (follow links from here) pretty well conceptually. His answer is Yes.

There have been thousands of posts and comments across the blogosphere since Nick took Krugman to task on the issue a couple of weeks ago, and Nick has been remarkably generous with his time in helping people understand his thinking. (A kudos also to Bob Murphy.) And it’s worth pointing out that Krugman hasn’t really responded to the core argument head-on. Feel free to follow the threads.

Here’s Nick’s model in brief, in my words:

Government borrowing/bond issuance today — considering only its costs, not the potential up/downsides of the associated spending — propagates incentives into the future, like waves are propagated when you throw a rock in a pool. Those incentives cause the old people in every period to consume more than the young people. In each future period, parents will eat some of their kids’ lunch.

Each generation consumes the same amount as they would have otherwise (because first you’re young, then you’re old, first you’re a child, then you’re a parent). But if government eventually has to tax to pay back the debt, the young people in that period are forced to consume less over their lifetimes, because they don’t get to eat their kids’ lunch.

(Nick acknowledges the point that Jamie Galbraith and others have been making for years: if the future GDP growth rate is higher than the interest rate, on average, the taxation never needs to happen, so the burden is never imposed.)

Here’s why I haven’t updated my priors much based on this thinking:

1. The wave model of propagated old/young bond-buying/spending patterns, extending to eternity, doesn’t seem plausible to me intuitively. It seems like the rock-in-the-pool ripples would probably flatten as they spread through time — maybe (?), as discussed in various posts and comment threads, because over the generations a certain percentage of parents bequeath their bonds to their children. That leaves us with “we’ll have to tax to pay for it eventually, so somebody will have to consume less,” which is pretty much where we started.

2. I wonder whether a real agent-based dynamic simulation modeling of Nick’s scenario, with continuous time and using differential equations, would give the same results. I’m not enough of a mathie to even guess, but I wouldn’t be surprised to see very different patterns and/or results.

3. The huge majority of government bonds are bought and traded not by people but by institutions (many of which — this seems significant — are licensed to create credit money and associated debt ex nihilo, and use government bonds as debt collateral in that process). Those institutions don’t have generations — birth/death, parents/children — and they don’t consume real goods (much). Again, my intuition tells me that these facts would bring complex dynamic interaction effects into play. While it might suffice to simplify by modeling things “as if” children were buying bonds from their parents, I don’t feel confident that that’s true.

4. The question Krugman was really asking, underlying his “is debt a future burden” locution, was: A. should we be taxing more or less? and B. should we be spending more or less? (Hence, should we be borrowing more or less?) Since I’d expect to see complex interaction effects from any of the four sources/uses choice combinations, isolating the question in this way seems like a questionable analytical technique. It’s kind of (not wanting to offend, can’t resist the word) petifogging. I’m not at all sure it provides useful information when divorced from the relevant context.

5. Semantics: assuming the model’s right, that we’re forcing a future generation of people to pay for their parents’ (our?) extra consumption — but not changing the amount that all future people will consume in toto — should we call that “a burden on future generations”? I’ll leave that to the philosophers.

I may (still) be displaying an inadequate understanding of the model in some points here, but I think there’s enough of merit above to discourage certainty — to question the ultimate utility of the model.

In short, if I was running a business of any size (yes: I have done so), with any decent amount of money on the line, I would 1. not give a huge amount of weight to the results of this model, and/or 2. be asking for a much more sophisticated analysis.

Which (#2) leaves us again where we were, wrestling with many/most of the big questions of growth macro.

So when Nick says “I thought we all had this debt burden stuff sorted out 30 years ago,” I agree. The answer was “maybe.” (Especially given the long-term historical reality of the Galbraithian scenario described above.) And in my mind it still is — with a little more weight on the “burden” side.