From TIPS to BAGS, or BANGS

Back in June I got into a game of blog tennis with my Hartwick colleague Jason Antrosio, including this from me and this from him . That was June, and I am way overdue. This may be something of a response.

Here's the gist of the story so far. Jason and I agree that there are various infrastructure projects that would be really good to undertake, because we've already got a foot in a world where a combination of climate change and lack of easy oil means we need to be burning less fossil fuel. So we have these projects where, if we spend money on them, we'll be a lot better off than if we don't spend the money.

Under normal circumstances, borrowing for such things is a no-brainer. If we don't spend, our economy will grow at 1%; if we do spend, the economy will grow at 2%. A faster-growing economy provides more output for the government to tax, and the bonds can easily be repaid. But there's a catch.


I look at the prospect of energy prices that stay high for years or decades, and at oil supplies that increase much slower than we're used to, and I see a global economy that has trouble growing. In that context, these necessary pieces of infrastructure could mean that the economy will only shrink by 1%, instead of shrinking by 2%. At some level, it's not too different from taking out a mortgage in the expectation of a 2% raise every year, and then finding that you're getting a 1% pay cut every year instead; what would have been a perfectly reasonable debt load in one case turns into a crushing burden in the other.

Of course, the the interest rate on a 30-year Treasury bond is still down near historic lows, meaning that the government can borrow money relatively cheaply, which in turn means that a given amount of borrowing right now at low interest rates translates to less of a crushing debt burden in the future than if we were borrowing at historically normal rates. (Jason addresses an implication of that here.)

And a bond is usually denominated in nominal terms: the borrower promises you a certain number of dollars in the future, not a certain amount of purchasing power. One of the risks with a bond is that the borrower promises to pay you $1,000 30 years from now, and then we get to that future and it turns out there's been more inflation than you expected, so when you get your $1,000, it's not worth as much as you'd thought it would be.

And that leads to two possible interpretations of the current low interest on US government bonds. The standard one is that players in the market are signalling that they don't expect much inflation over the long term. Lending someone money for 30 years at a nominal interest rate under 3% only makes sense if you don't think there's going to be much inflation over the intervening years.

But another interpretation is possible. What if bond traders share my gloomy assessment? (I doubt they do, but let's just explore the implications.) They expect the real economy to shrink. They expect the government to repay its loans by allowing significantly higher inflation than we've had in a while, maybe something up around 5%. In that case, lending money at 3% is roughly equivalent to losing 2% of your money every year. That sounds pretty bad, but if the economy is shrinking and most people are losing 5% or more every year, then losing 2% looks pretty good.

Again, I doubt that bond traders actually see things that way at the moment, but it does point in some pretty interesting directions.

First, it indicates a potential problem with something called TIPS, or Treasury Inflation Protected Securities. Remember, the big risk for someone buying a bond is that inflation turns out to be more than you expected. The interest rate on a TIPS bond is set by the change in the Consumer Price Index (our standard measure of inflation), plus a little extra. That "little extra" is set by the market when a new round of TIPS bonds is sold, and it represents the real interest that a TIPS practically guarantees you. If inflation turns out low, you'll get a slightly higher low interest rate; if inflation shoots up to a high level, you'll get a slightly higher high interest rate. So the TIPS removes the major risk facing the bond buyer.

And it costs the government very little to provide this service. If you assume that the economy will generally grow at, say, 2.5%, then the government incurs no new risk by selling a TIPS bond. If inflation is low, the government will collect relatively few dollars in taxes, but it won't have to pay many dollars to its TIPS bondholders. If the inflation goes up, the government will owe many dollars, but it will also be collecting a lot of dollars.

But there's one situation where a TIPS bond is a disaster for the government.

A normal bond guarantees you a certain number of dollars in the future. If inflation turns out lower than expected, then those dollars will buy you more than you'd thought they would, and the government will be handing over a larger piece of future wealth than it planned on; if inflation is higher than expected, the dollars you get back from your bond will buy less than you'd thought they would and the government will be handing over a smaller piece of future wealth than it planned on. In the extreme, if future wealth is actually shrinking rather than growing, the government can avoid being crushed by debt if it is willing to allow more inflation and thereby reduce the real wealth owed to bondholders below what those lenders had handed over in the first place.

A TIPS bond puts paid to this possibility. The government expects 2.5% growth. It sells some normal bonds that it thinks will require it to hand over, say, 2% of future GDP to the bondholders. If inflation goes down, that will turn out to be more than 2% of GDP, and if inflation goes up, the government's obligation will turn out to be less than 2% of GDP. With a TIPS, bond, the government know for sure that it will have to pay the bondholders 2% of GDP--not more, not less.

Knows for sure IF ... if GDP actually grows by 2.5%. If GDP grows slower, then the government's obligation under a TIPS, measured as a percent of GDP, goes up. There's no way to inflate your way out of the obligation, because the very structure of the bond protects the lender from the effects of inflation. If the economy actually shrinks, the TIPS become an unbreakable shackle.

The TIPS was a clever solution that protected both parties (lender and borrower) from uncertainty having to do with inflation, but it does nothing about uncertainty relating to underlying real growth. Fortunately, the mere act of phrasing the question that way suggests a solution.

TIPS protect you from inflation risk by defining the payout relative to inflation. Facing the risk of a shrinking GDP, we need a debt instrument that defines the payout relative to GDP. Something like Bonds Assuring a GDP Share, or BAGS. Better yet, we could sell BANGS, which would be Bonds Assuring a Nominal GDP Share. That way we cover both risks at once.

The government sells BANGS equivalent to 1% of this year's nominal GDP. The buyers--the people lending the government the money--are guaranteed 1% of nominal GDP in the future (the actual rate would presumably be determined through a BANGS auction, which sounds like a very entertaining spectacle). If nominal GDP has doubled, then the lenders will get back twice as many dollars as they paid in the first place. Maybe there was no inflation and the real economy doubled, in which case the owners of BANGS will be receiving a very substantial payout in real  terms; of course, incomes will tend to be higher overall. Or maybe there was lots of inflation and no real growth at all, in which case owners of BANGS merely get back what they put in, but they didn't lose anything to inflation.

And if the economy has shrunk, then BANGS owners get back less than they put in, but there's less to go around in general, and the claims of the bondholders aren't in a position to bankrupt the government.

It's worth thinking about how this system could be gamed. In a sense, a system of BANGS immunizes a government from having to worry about GDP: whether the economy grows or shrinks, the government's obligation to its lenders is equally bearable. And so, one might argue, the government will get lazy and not try to make the economy grow.

That's a natural thought for anyone who's had a couple of courses in game theory (and maybe for other people as well), but it's actually kinda nuts. When a government fails to produce a rising GDP, the political consequences are bad enough that it should be plenty motivated to do what it can for growth, even if it doesn't have to worry about the ability to pay back its bonds. It's far more likely that when degrowth comes, it will be due to forces beyond a government's control than to some sort of moral hazard problem where a government allowed an economy to shrink because it felt that its clever new bonds would allow it to be spared the consequences of not growing.

I need to mention a final issue, particularly in light of having spent the weekend at the biophysical economics meeting, a setting where nobody takes for granted that a growing GDP is inherently a good thing. I've written all of this consistent with an implicit assumption that GDP growth is good. It's true that with the way our economy is currently organized, a failure to increase the GDP is associated with undesirable phenomena such as increases in unemployment and poverty--and governments facing increased difficulty in repaying their debts. But GDP growth is also strongly associated with increases in emissions of carbon dioxide, and with negative environmental impacts in general.

So if we can address the problems of high unemployment, increasing poverty, and crushing government debts in a shrinking economy, there’s actually a lot to be said for a GDP that’s going down, not up. If BANGS are actually a solution to the debt problem, then we only have two out of three left to solve.

That should be no problem, right?

Alternative names:
DIPS: Domestic Income Protected Securities
BANDIT: Bonds Assuring a Nominal Domestic Income Take

Put your candidates in the comment section.