Sunday, March 06, 2011

Government bonds aren't net wealth...or are they?

In 1974, economist Robert Barro published a very famous paper entitled "Are government bonds net wealth?" You can read it here. The question it poses is a subtle one. Barro's basic argument is that no, government bonds are not actually wealth, since they will have to be paid off by future taxes; what looks like an "asset" is in fact also a "liability" of equal value. (This principle also goes by the name of "Ricardian equivalence".)The reason people treat government bonds as wealth, Barro argues, is because of a cognitive illusion - bonds are in your hand right now, while future taxes seem far away.

This is exactly the argument made by Tyler Cowen in a recent article in the New York Times:
We are fooling ourselves most of all. United States government debt in public hands is now more than $9 trillion, but most people still don’t realize what it will take to pay that off.

Here’s an example: Say that you have $20,000 in Treasury Bills. You probably believe that you own $20,000 in wealth. This will encourage you to spend and come up with ambitious plans. Yet someone — quite possibly you — will be taxed in the future to pay off the government debt. The $20,000 may be needed in order to do that.

The illusion is this: A government bond represents both a current asset and a future liability, yet for most people, those future tax payments feel less concrete and less real than the dollars they’re holding in a money market account...

In this case, the sorry truth is that our savings aren’t worth as much as many of us think, and a rude awakening is coming. One way or another, some of our savings will be taxed away to make good on governmental commitments, like future Medicare benefits, which we currently are framing as personal free lunches. 
It would have been nice of Cowen to mention Barro's name in connection with this idea. But be that as it may, I'm finding myself skeptical on a deep level about the basic argument.

Maybe government bonds really are net wealth.

Let's consider a related question: Are corporate bonds net wealth? Few people question that they are, and yet a similar logic applies. $20,000 in corporate bonds seems like an asset in your hands, but at the same time it is a $20,000 liability on the balance sheet of the company to which you lent the money. That company is going to have to pay that money back somehow, and it's going to do that at some point by charging higher markups on its products. You - or some consumer like you - is going to have to pay those higher markups until the $20,000 liability is paid off. These markups are analogous to taxes*. If people won't pay the markups, the company will have to default on its debt (just like a sovereign government might), and - Poof! - there goes the $20,000 asset.

Nevertheless, we commonly regard corporate bonds as representing some net wealth. The reason is that we assume that, before the debt is paid, the corporation is probably going to do something more productive with the $20,000 than you would have. They're going to invent the iPad, or build a building, or whatever, while you would have just stuck the $20,000 under your mattress and let it sit there. Hence, although the bonds might not represent a full $20,000 in net wealth, they represent some net wealth, because they represent an increased productive capacity for the human race.

How are government bonds any different? Well, government simply shuffles wealth around from one person to another - if government spending is 100% transfers - then governments do nothing productive with the money you lend them. In that case, government borrowing and spending simply shuffles money from the future to the present. Lots of libertarians seem to believe that all government spending is on transfers. Plenty of economic models - most importantly, Barro's - simply assume it right from the start. 

It seems pretty clear to me, however, that the existence of public goods renders that assumption invalid. If governments can do productive things that private companies can or won't do - create a functioning legal system, defend the country, fund basic research, or build a world-class highway system - then "Ricardian equivalence" goes right out the window. If government borrowing goes to fund spending on public goods that boost the tax base in the future, then that larger tax base can be used to pay back the debt...and the nation comes out ahead.

In other words, in a world with public goods, government bonds can be net wealth.

A quick Google shows that, unsurprisingly, I'm not the first blogger to point this out, but I think it bears repeating. When he claims that government borrowing is useless, Cowen is speaking not from scientific fact or logical necessity, but from a very shaky (and possibly ideologically driven) assumption. Barro's paper is more about the math of Ricardian equivalence, but his premise is much the same. And so both Cowen and Barro are awkwardly forced to conclude that people lend money to the government for fundamentally different reasons than they lend money to a private company (cognitive illusions vs. anticipation of a return on investment).

I prefer to give people a little more credit. Here's my counter-hypothesis: we expect a return on our government purchases of bonds because we expect the government to still be there in 2, 5, or 10 years. And we expect it to still be there because people will still need the public goods - roads, an army, and all the rest - that only governments provide. That doesn't mean we shouldn't be worried about public debt levels. But it does mean that the harsh fiscal austerity that many conservatives are now proposing might just be like sticking our money under a mattress.

Update: Discussing Cowen's column, Brad DeLong seems to agree that government is completely unproductive:
The key is that there is real wealth--factories, equipment, business organizations, and the profits that they generate--in back of the money market fund, while there is only the government's taxing power in back of government bonds.
DeLong seems to subscribe to this notion that government is essentially just a gang of armed people who take money from some people and give it to others. I strongly disagree, of course. Nonrival capital, and the institutional capital that allows government to effectively create more nonrival capital, are the real wealth in back of government bonds, even if the relationship is not exactly the same as for rival capital and corporate bonds.

* Of course, the choice to pay a markup or not is an individual choice, while the choice to pay taxes or not is made collectively by majority rule and representative democracy. However, this doesn't affect the basic argument.


  1. Benedict@Large8:28 AM

    By "paying back" federal debt, I assume we are talking about exchanging it at some point for US currency of some sort. And yet, US currency itself is issued as debt.

    The question of course then becomes how does one pay back debt with debt and ever get out of debt?

  2. Anonymous9:11 AM

    If your problem with debt is actually practical, rather than moralistic (even semantic,) than it's difficult to see why it matters that notes & fed deposits are technically liabilities. It's not as if the existence of dollar bills represents some kind of looming intertemporal burden.

  3. AndyfromTucson9:38 AM

    You may want to take a look at this link to the modern monetary theory perspective on deficit spending, where it is argued simple national income accounting dictates that "there can be no net savings of financial assets of the non-government sector without cumulative government deficit spending":

    In other words, some argue that net financial wealth is only possible with government debt.

  4. I love the irony that the Internet, an enormous wealth-creation engine created primarily by the government, has become such fertile ground for the likes of Cowen to spread the notion that all government spending is wasted.

  5. I would just like to point out that your characterization of Ricardian Equivalence is wrong - but in a way that gives the conservatives too much credit.

    A correct statement of Ricardian Equivalence is that the time-path of macroecon aggregates is independent of the time-path of the funding. Here, funding is the time-path of taxes. The problem is, of course, that RE says absolutely nothing about spending. As you say, Barro just assumes that gov't spending is thrown into the ocean.

    Only by disregarding the fact that spending is not in fact the same as taxes can conservative arguments actually make sense. What's worse, is that RE says that if the gov't spends, there is no benefit to putting off paying for the spending. In short, RE refutes, directly, every position of modern conservativism.

    So, conservatives bringing up RE in their arguments is the worst kind of dishonesty.

  6. BSE: you're right that some conservatives use Ricardian Equivalence to argue that fiscal policy is ineffective. But Barro's original paper, like Cowen's NYT article, allows that Ricardian Equivalence can fail due to cognitive illusions. The theory says that "either fiscal policy is ineffectual, or people are irrational."

  7. Anonymous10:03 PM

    "skeptical on a deep level"? So you should be. RE is the biggest crock of s--t in economics....well if it's not, it's certainly in the semi finals....