Should our government spend (and borrow) more at negative real interest rates?

I have read so many posts saying yes.  But is it so obvious?  Let’s put aside the stimulus argument, I’d like to focus on the rate of return argument alone.

Let’s say I could borrow money at negative two percent real, but my seven cousins, three of whom are crazy, would get together and decide how to spend it.  I would get a vote too and they would agree to spend it on me.  I would have to pay it back.

I say no.

Many of the infra-marginal federal dollars are allocated by formula, such as with Social Security, and the cousins don’t have such a big say in the matter.  I am grateful for that.  But is it possible that the hypothetical new federal spending might be controlled by the cousins?  And what if four of them are crazy?

How should I feel about the exorbitant cost overruns on California high speed rail?  You know, the line connecting Fresno and Bakersfield?  That wasn’t even the crazy cousins at work.  (Or was it?)

I absolutely do not see this one as a no-brainer.  By using the “should” language in your thought experiment you can take away the crazy cousins, at least hypothetically.  But in the real world they are still there, and the non-crazy cousins screw up pretty often too.

How close in your family would the spending decisions have to get before you would accept a deal like that outlined above?  How many people would turn it down, even with their spouse in charge of spending the money?


There are things we have to do, such as repair infrastructure. Do we do it today, when everything is on sale, or later, when things will be much more expensive?

Except it's tempting to spend cheap money on your own favorite-but-not-vital projects now; later on, when money is more expensive, you will still be able to get more money to fund vital projects, because they're vital.

Yeah, color me cynical.

If you presume "favorite but non vital", you are deciding it with your choice of words, not logic.

How would you respond to cost/benefit analysis to select project?

- money is fungible; can you do a cost/benefit analysis of the whole government spending? If not, the analysis may be an argument to finance one project - in absence of other government spending.
- if you did cost/benefit analysis of all projects (which I would say a liberal/government bureaucrat would say they are doing anyway), a lot of cost/benefit analysis would be significantly flawed (public choice). The response to a cost/benefit analysis would be: it is very likely significantly flawed.

"Vital" is the standard? How about beneficial instead?

Is a new tunnel under the Hudson "vital?" Life will go on without it. But building it is certainly beneficial, and when the federal dollars involved are free, why not?

I'm using words to express logic. If you'd like to substitute "popular" or "poltically defensible" or "favorable cost/benefit ratio" instead of "vital", feel free.

Even if it's "free money", you have to limited amounts of it and therefore opportunity costs.

My point is that Project A may have a benefit of +4 to the politician / bureaucrat and -1 to the citizenry, while B has +1 and +3. When you've got "free money" you fund A, and when times are tough you whine to the citizenry that unless they pony up, you can't fund B.

Sooo....ummm...what have the crazy motherbuggers been doing with the money they already had?

Iraq and Afghanistan, for starters

" let’s put aside the stimulus argument..."

I know, let's have a thought experiment, but restrict the most important part of the argument! That makes so much sense!

Ok, let's instead agree the multiplier on government spending in a highly indebted,otherwise pretty well developed economy is no greater than 1. Now we can incorporate the stimulus argument.

If you assume something that leads to your conclusion, why even make the argument? Is it true that the mulitplier is no greater than 1? Support?

I think that was a joke...

I believe Barro did studies recently concluding that the multiplier was always below 1. That war which had the greatest multiplier only reached .8. Further I think logic dictates the the multiplier will always be below one - unless the government can tax and spend with 100% efficiency - in which case it reaches one - barely.

Does everybody need to go re-read "The Broken Window" ?

dhlii, did you read Barro's op ed and the responses from other economists? He starts with a fiscal modifier of 0, arguing that at full employment, government spending has less effect than private spending. Assume full employment, don't do any empirical work, and write an op ed. By the way, as support, he uses a period of price controls during WWII for his argument. Good grief. Remember the conclusion, and not the support for the argument and how it was criticized.

dhlii, Barro starts with his 0 arguing that defense spending is of no value and deadweight. As for non-defense spending, he says: "Multipliers for non-defense purchases cannot be reliably estimated because of the lack of good instruments."

dhlii, Menzie Chin has collected the literature on fiscal mulipliers, if you are interested:

If the government should not spend and borrow at negative rates, why should the government spend and borrow at all (outside social security etc)? That is, why would you want to make any decisions together with your crazy cousins?

Agreement with your argument is an agreement with the position against any discretionary spending, even at negative rates. It's an argument for a substantial reduction in government spending.

That isn't necessarily true, because the argument is in response to an argument for increased government spending. It is possible to concede the existing level and still oppose the increment for the reasons stated.
Of course, one could also conclude that a substantial reduction in discretionary spending would create more value for citizens by eliminating inefficient aspects of such spending.

The argument is in response, but the argument is general enough to include maintaining the same level of spending.

Anyhow it reminds of the distinction between buying a stock you don't have and keeping a stock you do. These decisions are identical.

I think you are missing Tyler's point.

There are some things government just needs to do. You spend to get these things done, even though you have to use the crazy-cousin system. Because there is no alternative, efficiency is not a valid argument against this type of spending.

However, this is the argument before us: borrowing money is free or slightly better than free; therefore, independent of any specific need, we should borrow just because it's such a good deal. Inefficiency in the spending defeats this argument, because it cancels out the benefit of "free" borrowing.

Tyler's argument is not against any spending and borrowing, only against superfluous spending and borrowing because it's such a good deal. Of course, then it becomes a debate about what is needed, but that is where the debate should be.

I have to disagree. If you can defeat the Nazis for a million instead of a trillion, it matters.

I think you would first have to ask: what do I expect future interest rates to look like, and do I expect to roll this debt over? If you think they will be high and you will not roll over, it's much more attractive. If they are going to fall even further into negative territory, or you are rolling over...

And, lest we forget, U.S. debt is heavily short-term, so we're already looking at serious problems if the bond markets suddenly decide the interest rates on that debt are too low, let alone if we take on trillions more. So expecting higher interest rates in the future doesn't necessarily mean we want to take on more debt now (since we presumably aren't planning to pay it off).

The "let's build infrastructure!" arguments don't hold up very well either. We already have lots of infrastructure; more is not necessarily an efficient use of money. Unless you believe that infrastructure's marginal utility does not decrease no matter how much of it you build, it's very reasonable to ask if we're already spending too much on it -- esp. when considering those high-speed train boondoggles.

@TallDave I get that it's hard to say whether new infrastructure investments are a new money (although I would also emphasize that, unfortunately, government spending is the only way we get infrastructure in America and there are places like Los Angeles that are clearly deficient in infrastructure).

But I don't get the arguments against maintaining existing infrastructure. Not maintaining existing infrastructure basically defers spending and possibly increases the total amount ultimately due. It sounds like we can do it on the cheap now, or expensively later.

I don't have strong feelings about fiscal stimulus or anything, just if we're going to do it at some point now seems like a pretty good time.

The infrastructure needs to be maintained, but it's not clear that we should accelerate the schedule of maintenance beyond what needs to happen regardless of interest rates. There are two reasons for this. First, we may not be able to in fact accelerate infrastructure spending. ARRA tried very hard to do that, but states drastically cut back their own infrastructure spending in response to federal infrastructure dollars, resulting in little if any net increase of spending. Second, if we can in fact drastically raise spending, we may end up paying more than we need to for the maintenance because we bid up the prices of the qualified firms. Tyler was making the point yesterday that much government spending is not going to help unemployed persons, but is rather chasing after skilled and busy workers/firms. They may just respond to increased government demand by charging more, causing any savings from cheap borrowing to vanish.

There's a factual issue here, and that's that maintenance budgets have been hit extra hard during the recession because maintenance delays are one of the first things municipalities and states do to fix sudden budget deficits. The backlog exists (to the tune of a few hundred million, in the case of WMATA) already, so it's not like you would be be accelerating the schedule ahead of what the engineers want. You're just catching up.

John: Well, then you get into questions of efficient allocation resources. Why does it cost 25x as much to build a bridge, in real dollars, since Tappan Zee was built? Does LA need more roads, or are the existing roads mispriced to consumers (i.e. are they overconsumed because of the implicit subsidy)? I've heard there are proposals to build new express toll lanes on top of existing lanes. Such proposals tend to be defeated politically because as much as people like to avoid traffic, it makes the people who don't wish to pay for that privilege upset.

I'd dispute your claim that we already have a lots of it. Much of what we have is old and in decay. Out my window I can see the bridge over 35W that collapsed a few years ago in Minneapolis. There are many others like it that could stand to be overhauled.

I take your point about rolling over, but it seems that money is currently cheap, the bridges need fixing, and the marginal utility is similar to when they were built in the first place. Why not?

Adam, I thought it was well understood that that particular bridge collapsed due to design flaws, not because of being old and in decay. TallDave was right to point to the likelihood that interest rates will go up, and any debt incurred now adds to a total that is unlikely to be paid before then. If the answer is that the Fed is just going to keep rates low no matter what, then I'm not sure how that would be different from just printing money to pay bills. So that's why not (i.e. why not to borrow more because rates are cheap).

But the productive capital value of Bridge 9340 was just as high as that of the I-35W Saint Anthony Falls Bridge you are looking at today, up until August 1, 2007, 6:05 p.m.

Just because 20,000 bridges are getting much closer to bridge 9340 in 2007 doesn't mean we should buy 20,000 I-35W Saint Anthony Falls Bridge which are designed for the future and cost far less because of the weak economy.

Besides, life didn't end when bridge 9340 failed - it created untold numbers of jobs to drive the detours in the increased congestion.

So, we haven't been maintaining infrastructure...because interest rates were too high?

Or...because the governments are a bunch of crazy motherbuggers?

TallDave, I wish you hadn't made the point about infrastructure, because everyone is commenting on that instead of your excellent main point that if we borrow more without planning to pay it off we are going to have to roll it over anyway under those future higher expect interest rates...

This is a fundamental point and I am astonished that it seems to be ignored by those arguing for more spending now because interest rates will be higher in the future...

That's right and I was going to make the same point and I'll re-emphasize it here for fun.

It is built into their argument that the money is "free" BECAUSE interest rates will be higher in the future.

So, you have to get a higher ROI with these projects than we have been getting, it doesn't justify a lower ROI just because the money is "free" (it's not free, we have to pay it back).

All projects with positive NPV should be funded. The problem is there are so few. Most govt spending is for consumption anyway - Social Security, Medicare, etc. Other spending is for regulatory or administrative purposes - overhead basically, no incremental NPV projects there. Education spending has virtually no payoff - it's been increasing for years while the economy goes nowhere. Military, law enforcement, TSA spending, are costs you have to pay but there is little in the way of incremental projects in those areas that are positive NPV. Public goods are great,l but it's hard to value them, so hard to find NPV in those. This list probably accounts for >80% of all govt spending.
Most "infrastructure" projects turn out to be repair projects. They should be done when repairs are due, regardless of economic or fiscal effects. They are NPV positive compared to the alternative of letting things get worse, but they don't increase economic activity, they just keep the status quo. And repairing things ahead of time to increase activity is just a timing trick that should be avoided. A lot of projects obviously tend to go over budget. At the end of the day there are very few NPV positive projects that govt should borrow to fund.

"Education spending has virtually no payof"

That's absurd.

"Most “infrastructure” projects turn out to be repair projects. They should be done when repairs are due, regardless of economic or fiscal effects"

Sure, but they don't get done. So why not do them when the economic times are bad and money is cheap?

The marginal value of spending more on education appears to be low, to put it mildly.

It's really somewhat insane that education spending has skyrocketed even as the price of information dispersal has plummeted. This is subsidy run amok.

The thing about negative real interest rates is that they increase the number of projects that are NPV positive. Compared to a world with higher interest rates, there must be some projects on the margin that *should* be funded now that interest rates are lower. Those are the projects we should indisputably be borrowing cheaply to fund.

Tyler's argument boils down to claiming that the people in charge of appropriations are unable to correctly identify those projects, and moreover that they'll do such a bad job that they will instead identify projects that, in sum, have negative NPV even at low rates.

Mark, I think it might be more accurate to say that education spending by government, on an annual basis, has already long since entered the realm of decreasing returns. Also, I agree in principle that repairs whould be done when they are due, but here we encounter one of the problems with political allocation of resources. There is more political (i.e., vote-buying) bang for the buck when spent on highly visible new projects. Eventually those new projects become old infrastructure needing repair and replacement, but the next generation of pols wants to buy its own votes with new projects. So there is some combination of out-of-control spending and neglect of needed maintenance.

All projects with positive NPV should be funded

Even if you crowd out projects with higher NPV?

High speed rail is horribly expensive. It only appears "economical" when you lie about the costs. Part of what makes it so expensive is that there can't be "at grade" crossings - every crossing is a bridge of some sort (overpass for the road, overpass for the rail line). The reason that at-grade crossings are a no-no is a simple matter of trig: the curvature of the earth means that the horizon is closer to the train driver than the distance it takes to stop, so by the time that the engineer identifies a threat, and hits the brakes, it is too late to stop before hitting the obstruction.

For some details on existing HSR:

Distance to horizon:

While locomotives in the US tend to run about 15' tall, the eyeballs of the drivers are about 2' lower.

Grossly underestimating the final costs is very common in large infrastructure projects. Whether those be "the big dig" or California's HSR, or some of what is going on here in Denver.

Your crazy cousins probably wouldn't have bought that expensive thingy when it is priced at $X. But they'll think it is a great deal if it is priced at $X/3 and the folks "selling" it will depend on the sunk-cost fallacy to keep those cousins throwing good money after bad.

Over the horizon radar, satellite tracking, video at all at grade crossings etc. Seems to me there are plenty of technological fixes for this besides more expensive bridges and tunnels.


Most Interstates have no at-grade intersections.

All interstates. It's in the specifications.

According to Krugman and Keynsians it doesn't matter if the cousins are crazy...even if they just pay to dig up holes and refill them it will help make us all better off.

Hey! Don't forget Friedman's helicopters!


Digging and refilling holes could be better than the kinds of things the crazy cousins come up with. At least that project doesn't consume any capital.

Can we just borrow and use the money to pay off old debt? At negative interest rates, it doesn't seem like there's a credible argument against that.

Holy crap. Somebody just had a useful idea. The treasury must be doing this already, but then again it's sometimes surprising what obvious sensible things don't happen.

This is what I would do, would that my credit card company offered me this deal.

Unless I'm mistaken, aren't Treasury bonds/notes non-callable? In other words, they can't be retired early...

They are however buyable.

If the market is working correctly there is little if any difference in the secondary market from the primary market. So if the fed would be able to issue debt at negative real interest rates, the it would have to purchase back its own debt at negative interest rates.

Would be great if T-Bills were callable!


"How close in your family would the spending decisions have to get before you would accept a deal like that outlined above? How many people would turn it down, even with their spouse in charge of spending the money?"

Hard to figger how intelligent people keep using this "family spending/government spending" analogy. I am not an academically qualified economist but isn't this a pretty poor comparison?

I would think a "giant corporation spending/government spending" analogy would be much closer. And, although there are plenty of crazy cousins in giant corporations, I have a feeling that borrowing at low/negative interest rates might be seen as a good idea for giant corporations even by some conservatives.


Corporate spending is usually on something that will return a profit, home spending is consumption. Gov't spending is much more like the latter.

I would think building highways, meteorological forecasting systems, Air Traffic Control systems, postal systems, universities, basic research science foundations ad infinitum, could be compared to corporate spending. There are lots of things that governments do that provide infrastructure and a stable environment that enhances GDP growth and therefore increases their revenue.

Also a family budget can often safely carry a debt load of 200% or more of annual income. A sovereign state is considered to be in trouble when debt gets to 90% of GDP. And how many families take out debt in the form of bonds like corporations and governments.


The analog between a family's income and a government's is tax revenue, not gross economic output. 90% of GDP would be ~350% of annual revenue in the US.

Well sovereign debt load is usually expressed as a percentage of National Income and that is why I used GDP but my whole point was the "family spending/government spending analogy does not make as much sense as a "giant corporation spending/government spending" analogy. I mean - except for maybe the Sopranos - how many families print their own money.

We could consider a giant corporation's total revenue as like a government's GDP and a corporation's profits as like a government's tax revenue. So how serious a problem would debt be if a corporation were paying 6% of profits servicing its debt like the USG. Even in bad times I would think they would borrow if rates were low to save trained personnel, production capacity, etc. so they could make the most of it when the economy picked up.

The USG can borrow money for 30 years at real rates around 1%. It seems to me by borrowing less they are saying that 30 year investments cannot return more than 1% and the rest of the country (world?) is following their lead. Can this really be the way out of the slump?


You can't analogize a family budget and the Federal one. Families can't print money.

No. Most corporate spending is on supporting the things that return a profit. Call it infrastructure (engineering), marketing (and pr), and human services (it, hr,call centers) . The sales guys are actually fairly cheap, but they are the only ones who make a corporation money. This is exactly like the government, except that a government doesn't have sales outside of bonds. Perhaps more of government should directly support the trust in bonds? That actually sounds like a terrible idea.

When buying for home you never invest? To mow your lawn, you go out and rent a mower because you don't invest at home. When deciding to go somewhere that requires a car, you call a taxis because you never invest. You live in a hotel to avoid investing?

The past three decades I've seen the message to individuals to be to focus on consumption and to stop investing.

Don't invest in a house, use it as an ATM.

Don't buy a car to get you to and fro for the lowest cost, buy a car as a luxury.

Don't ever ask about energy efficiency because government will drive down the price of oil by drill baby drill so using 3000 gallons of heating oil a year is always going to be cheap.

Don't buy sturdy durable clothing because you should discard clothing after a few weeks to be stylish.

And when it comes to government spending, it seems the idea is to pay people for not working because replacing bridges at risk of falling down is just wrong until the bridge falls and kills people to generate the press coverage to get earmarked pork passed in Congress. NH and Maine DOT officials are idiots who don't get it - they are closing bridges before they fall which ensures years of delay before money funds replacements. The Pawlenty administration was much more effective, it obtained the Federal funds in less than 4 months by operating the bridge to failure, and completing the new bridge 12 years ahead of schedule.

Loren, Bingo.

The family analogy can also be reframed as a family taking out a mortgage to buy a house knowing that their earning potential will be their to sustain that decision.

I think the corporation analogy is a good one. I like it when people say: we have these unfunded liabilities. Yeah, corporations have them too: their called bonds, and so long as the corporation can continue being profitable and raising revenue, those IBM bonds will be paid off. The market would tell you if they are a bad bet too: you wouldn't buy them or their rate of interest would have to be very high if it were a bad bet.

US debt is low. The market is telling you something.

The family analogy is taking a line of equity to pay for a vacation, because we don't earn enough to pay for it.

US debt has been downgraded. The market is telling you something.

S&P is not a market. S&P is a private actor with diverse incentives.

Let's say I tell you gold is worth no more than $10 an ounce, but the market says otherwise. Who are you going to believe?

Until we end noise like Davis-Bacon, spending on anything new is retarded.

Baumol says that the rate that the government can borrow at is irrelevant:

First of all, I want to thank Tanguarena for teaching me something I didn't previously know about railroad crossings. Learn something new every day...

To address the topic at hand, you borrow at a certain rate (real or otherwise) because you think you can get a higher return on that money (real or otherwise.) If you think you can arbitrage the yield curve by using the proceeds to buy back "undervalued" debt (relative to your borrowing costs), then borrow away. Just make sure you have the math correct!

For spending projects, the "return" part is very hard to figure out, since there are usually no immediately measurable benefits. How does the country get a return on welfare spending? Fewer food riots? Less crime? The ability to support children who will later grow up to be productive and taxpaying citizens? What about defense spending? Will the next aircraft prevent 10% more damage to the country (or its interests)?

Who knows? It's all so politicized. Then again, it's tough to tell old people with diabetes to suck it up and deal with the blindness, or to throw some single mother out in the street.

Spend money like a good chef salts food.

Hi Tyler--you keep telling us how much more clout you want scientists to have to mystically lift us out of the great stagnation. Here's your chance! ASCE tells us we must spend approximately $2.2 trillion. I can see 4 possibilities, but tell me if I've missed some:
(1) you've decided that we should not heed them, in which case your argument about scientists rings hollow.
(2) your argument doesn't apply to ASCE in particular for some reason. why not?
(3) you think there's a non-government solution to the problem. What is it?
(4) you agree with ASCE, but think the government is incapable of dealing with it, despite the history of the interstate highway system. The "true" cost after factoring in waste, inefficiency, blablabla, is much higher than ASCE thinks, making it ultimately "not worth it." So what happens next? How does our country compete with the rest of the world? While we're at it, how has so much of the rest of the world managed to pull it off so much better than us?

Don't be so disingenuous. You clearly do not understand the benefit scientists with political capital can play. Mysticism would be if Tyler wanted priests to have more clout. No thank you.

Better question:
How many of you would you would borrow a lot of money if you could borrow at the terms and rate the Fed Gov. can?

I think many would say no.

But I *can* borrow a lot of money at close to Fed rates, and I am saying, 'No'. I'm refinancing my mortgage at 3.25 for 15 years while a 20 year treasury yields are at 3.1. My house has a lot of equity that I could take as cash out -- but I'm not doing it.

Well, if I, like the federal government, knew that:
1) My annual income was extremely unlikely to fall by more than 5-10% (and that in a terrible recession), so I didn't need to worry about sudden change in ability to pay due to changing my job, and
2) If I really had to, I could print my own money to pay it back

Then as long as I thought I could invest in something real and/or productive, of course I would borrow the money.

This gets to the core of the issue: the interest rates are irrelevant -- except that if they are particularly favorable, it may encourage you to fund projects with lower ROI -- it all depends on how good of an investment the project it is, and then by extension, the competency of those deciding on those projects: in our case, presidents, senators and congressmen.

I think at core, no one in the world really believes the "all spending is equal" thing. Even if you think we should spend more as stimulus, no one is truly ambivalent about the destination -- an anti-war pro-spending person would not advocate the "stimulus" of invading Syria, no matter what Krugman implies. Every pro-spending person has a good idea that the money should be spent on, and they don't think they're crazy cousins.

Every single person advocating spending increases on project X (and actually, spending cuts as well) should be legally required to say "Assuming we replace a certain 535 people with those who will agree to vote for project X..." before giving their advice.

But the thing is, we already know how our own seven (or 535...) crazy cousins will vote to spend the money: because we have their voting records. They will vote for 1) more wars 2) projects in their districts/states, regardless of merit 3) re-allocation of wealth to their constituencies (old people, campaign donors). If you find yourself advocating increased spending, you are in favor of those things -- unless you also have a Manchurian candidate-esque plan to replace 51% of those crazy cousins.

Tyler, let us go back to the classroom. We all have an idea of how we would answer a question about the net benefit of borrowing an additional dollar to spend. If I borrow to buy an asset, the answer will depend on the benefit from using the asset and the marginal cost of borrowing. If I borrow to have a good weekend, the answer will also depend on the benefit of whatever I'm going to do over the weekend and the marginal cost of borrowing. The difference between the two cases is that in the first one I'm expecting to pay the principal of the debt back with the asset I'm planning to buy (or with the depreciation fund accumulated during the time I use the asset), whereas in the second I will have to pay it back with whatever funds I'm able to get from some other source (to simplify, either savings accumulated in the past or future earnings from other activities). In other words, when I borrow for consumption I have to take into account any cost associated with paying the principal back (or if I don't pay it back, with default). Indeed, we often assume that borrowing to consume amounts to optimize consumption over time given the income flow we expect to earn over our life and therefore we assume that the cost associated with paying the principal back is zero --at least, as long as we take the income constraint into account.

Now let us go back to government borrowing. If it borrows to buy assets, the key issue is how large the benefit from each asset will be --and we know that there are many assets from the which the government fails to earn directly or indirectly a benefit large enough to justify EVEN a negative real interest rate. If it borrows to redistribute income, sooner or later the key issue will become the cost associated with paying the principal back --in other words, whatever the benefit from increasing the consumption of some group of residents that may have justified initially to pay interest on the debt, the benefit will have to be large enough to also justify the cost associated with additional taxation or with reducing other government activities.

Indeed, if the marginal cost of funding the government were zero or negative, the rational choice would be first to re-finance the outstanding debt so the average cost of funding it is reduced at least to zero. Thus, to answer Tyler's question we have to proceed in two parts, first let us see how far the government can go in re-financing its outstanding debt to reduce its average cost, and second let us see how large the marginal cost associated with paying the principal back is and how it increases as the government continues to borrow to buy bad assets and to redistribute consumption.

How about we rephrase the proposal. The government already does spend lots of money. When real interest rates are negative, we should finance that spending with borrowing rather than with taxes.

Based on the idea that the tax cuts have put so much money in people's pockets they won't spend on productive capital in any form, not for VC funds for a new startup that will create productive capital, not for stock in a new IPO for a new company buying productive capital, and not in loans to small and medium firms to pay for buying productive capital, these tax cut dollars can only bid for US Treasuries.

In other words, the market has dictated the best place to spend the tax cut dollars is in government spending because no other spending increases the economic welfare, efficiency, etc as well as government spending.

If taxes are hiked, the dollars get spent by government without any change in spending, but the future requirement to repay the debt is eliminated.

So the spending is a given which can't be altered, and the only question is do we raise taxes or increase borrowing? Well, I guess that is one way to frame the problem. But other people might think of, say, holding the taxes and borrowing constant and adjusting the government spending.

"When real interest rates are negative, we should finance that spending with borrowing rather than with taxes."

Lets rephrase that:

"When real interest rates are negative, we should finance that spending with future taxes rather than with current taxes."

And then you see why its a non-starter.

Prof Cowen, I don't believe your analogy gets at the crux of the issue.

Your situation assumes a prior commitment to borrow a certain amount and then let crazy people vote on how it's spent, with you bearing the proportionate (or disproportionately high) obligation to pay it back. And it's true that many policy prescriptions ("theory suggests a trillion dollar stimulus is needed now") fit that model.

But in the real world we can offer specific programs beforehand. Infrastructure spending, payroll tax cuts, etc. So it's closer to a situation where you can borrow money while earmarking it for a specific use. So you and your crazy cousins can get together and borrow a couple grand to retrofit your houses for earthquakes, buy health insurance for everyone, etc.

If President Obama came to you and offered a specific list of projects (more specific than things like "high speed rail") and offered you a line item veto over the projects to be funded via borrowing, I'm guessing you'd be much more willing to jump in, assuming the existence of appropriate enforcement and follow through.

A second thought: does the craziness of the cousins really play an important role in your model? The issue lies more in the reality that you'll pay disproportionately in the future for the borrowing made now, as your future income will likely continue to outpace theirs, and lenders don't care who of you pays them their money. So would a more regressive tax structure, with guaranteed tax levels into the future, weaken the strength of your argument?

If only Tyler would carry his thoughts one step forward and make the point I have been making, and Shiller recently did:

taxes must be hiked to invest in infrastructure and human capital, and oh by the way, create jobs.

The reason taxes were hiked in the early 30s was the people with the earnings who would have to pay the taxes weren't spending the money on productive capital investment when the nation had huge productive capital investment needs.

What strikes me as the trend over the past three decades is for less and less investment in productive capital.

Less is spent on R&D, especially the basic research by both public and private.

Less is spent on transportation systems with the result being declining levels of transportation service at all levels.

Less is spent on water infrastructure of all sorts so we have either too little water for basic needs, too much water, too polluted water, costly pure water wasted when cheaper gray water would do, yet water demands continue to grow with only finger pointing and law suits instead of investing in appropriate water related productive capital.

The fact the US government can borrow so cheaply proves those with wealth and income can pay far higher taxes to fund productive capital investments in the US, but without those investments, no private productive capital investment offers a return higher than 0-2% for a high risk, so a -0.5% is net-net a more profitable investment.

If you wanted to build a foundry-metal working operation in Pennsylvania or Ohio where those operations were once common, you would find it nearly impossible to ship your product - better to build your factory in Asia where the government makes transport infrastructure for global trade a priority.

How about a revised question: should our government not reduce spending at all and instead borrow more at negative real rates? That seems like a no brainer to me.

As does Treasury doing what it can to restructure its financing to be as heavy on the 5-7 year window of negative real rates.

Beyond that, I agree that you have to know what you are spending for. But that can be done in the legislation.

The question, should our government borrow more because of the negative real rates, has been answered by those who oppose deficits: we will not be issuing debt, and borrowing more, even though there are negative real rates.

Congress will be meeting in October to cut spending by $1.2 trillion over ten years. Sure hope they don't front load it, but they are passing up an opportunity to borrow more at lower rates. Maybe we'll have even lower negative real rates as people seek out more security by purchasing riskless assets which will be in shorter supply.

Since the crazy cousins are Republicans and are going to say no to anything you propose anyway, I don't know why this question is even interesting if we're taking it past a thought experiment.

Then again, your cousin might think you are the crazy one.

If he is the one waving the gun around, threatening to kill your sister if you don't give him money, it's probably him.

In the housing bubble and its aftermath, a lot of people dug themselves a big hole by borrowing more than they could afford at low "teaser" rates that later went up. How is it different for the pols to plunge us deeper into debt now?

One should also note that is the money is spent in such a way that it boosts economic growth, inflation will probably increase somewhat, meaning the negative rate debt becomes more profitable in real terms, but more problematic to roll over. There's probably an optimal strategy that can be worked out from this. Obviously, if we could significantly boost growth for 5-7 years and then cut back on spending when the economy is healthier, that would be ideal. Which is just to say that if Keynesian stimulus is a good idea, it's a really good idea when you have negative real rates.

Pat, You're right, but we just committed ourselves to austerity with a debt reduction commitment of $1.2 trillion beginning in October.


Over ten years with the bulk of the cutting happening (as always) late in the the term.

At a time we're overspending >1.2 trillion/YEAR. Austerity??

It's fiscal drag if you cut spending from current levels. I'm in favor of raising taxes and cutting spending later.

None of this would lead me to favor a massive increase in direct federal spending. I could be persuaded to consider the merits of a well-chartered infrastructure bank. That is, an entity that lent federal dollars to some combination of state/local governments and private concerns at market rates and performed the necessary due dilligence. Yes, you might still have some of the corporatist rent-seeking reasserting itself at the sub-federal level, but in theory that is easier to control than a big federal free-for-all like the ARRA was.

What disturbs me most is contemplating the underlying socio-political framework that can result in a situation like this. Negative real interest rates? Essentially, people are begging the government to steal their money.

It bothers me that people could say, "We're crazy not to take the money!" when you think about what this actually means.

In the abstract, I would agree. Bureaucracies (in the private and public sector) are inefficient. Concerns about unsustainable long term deficits should make us pause before mindless borrowing. And yet, there are plenty of concrete examples in which public returns (due to various market imperfections) far exceed private returns. Even factoring in government inefficiencies, the return on additional spending may exceed the costs. Some possible examples might be more funding for tech start ups or wage insurance for the long term unemployed or principal forgiveness for the deeply underwater households. Of course, policies which are economically desirable may not be politically feasible.


1 ."more funding for tech start ups"--I can see funding early stage, basically basic research through universities, but after that, what is the market imperfection that favors government as an investor?

2. "principal forgiveness for the deeply underwater households"--what is this market imperfection that requires the government to do this? I could see the government giving a different interest rate in exchange, for example, the homeowner either sharing in the gain on resale or paying tax on the gain. But, why naked forgiveness?

Agree on 1. Go for basic research. Sure the marginal product will probably be less productive, but you still might fund a gem. I think you could do this by pushing the young researcher awards at NSF, NIH, etc. Give it to people who are not established enough to 'win' at the usual grant writing game.

On 2, suppose for a moment that a market failure, like allowing no-doc, low-collateral loans, actually led to some consumers taking on mortgage debt that they simply could not. Now they did sign on the dotted line, but should these households be held accountable to the almost predatory products they were peddled. Research by Annamaria Lusardi and others shows that most people don't understand the concept of compound interest. What does that mean in the mortgage market? I think it means you only understand monthly payments...and that sets you up for accepting all kinds of mortgage contracts that are very bad for you. Yes, it kind of stinks to bail out people who took on too much debt. And yet, their insistence at paying such mortgages (as long as they can) seems to be mucking stuff up in the economy. So I would push 'naked forgiveness' to make up for a prior regulatory mistake. You don't need to respond and tell me how politically (or personally) unpopular this is. And all the future moral hazard issues it could raise (though the bailout of AIG had some of those features too...) I know this one is a non-starter. I was just using it as an extreme example...


I'm open to number 2 based partly on your argument, but I don't know how people would support this on equity grounds, unless there were a justification (as you offered) AND some contingent or actual loss by the person who received the benefit so that the polity could rationalize they did something that helped the economy (read: themselves) while possibly helping others (in a recession, fearful people are none too altruistic).

So, if you framed it as: 1. the government is going to SAVE money by restructuring loans (as a bank does), will take a hit ( as a bank does) to get this off their books (as a bank does); 2. the debtor will continue to pay, but with lower rates and over a longer time (at rates the government borrows?); and 3. they will pay it back via sharing in any appreciation if they sell the property or paying capital gains taxes (you could just say Paying taxes) on any future gain.

Also, re your report, have you contacted Thaler yet for some questions or critiques.

People, I've found, are not too altruistic, unless you can make them see it is in their own interest (which, of course, is a contradiction).

The negative real interest rate is just a brief market aberration, I suspect it will not last long.

The headline is a false choice: maybe we don't need to spend more, but spend differently to stimulate the economy more.


"So are we stuck between a fiscal policy rock and an economic hard place? Not necessarily. It’s possible to stimulate the economy further, now, without increasing the 10-year deficit. It would mean steering funds away from spending and tax cuts that offer low “bang for the buck” economic stimulus and toward higher “bang” policies.

Then the deficit would be no higher but the economy would improve. That growth, in turn, would help brighten our debt outlook, as our economy would be better able to keep pace with our still-growing debt.

What shifts would be needed? The nonpartisan Congressional Budget Office lists payroll tax cuts and aid to state and local governments high on its list of fiscal policies designed to stimulate demand. Care to guess what falls dead last on that list? Extended income tax cuts.

So why not let the Bush tax cuts (at a cost of $2.5 trillion over 10 years) expire, as scheduled, at the end of 2012? At the very least, we should insist that any portion of those tax cuts we wish to extend should be paid for.

Healing both of our economy’s ailments isn’t hard as long as we open our eyes to all options. Round 2 of the debt-limit deal gives us another chance to get it right."

I totally agree "steering funds away from spending and tax cuts that offer low “bang for the buck” economic stimulus and toward higher “bang” policies." Even the most effective government agencies have gross inefficiencies that could be addressed. Another idea would be to use new funds to make process improvements at government and quasi-government agencies, like electronic, portable medical records for everyone. I disagree that payroll tax cuts should be high on the list of stimulative policies, but a lot depends on your assumptions and your benchmarks. Just because a policy like payroll tax cuts is better than other policies in an ordinal sense does not mean it is 'above the bar' in a cardinal sense. PS love the blog.


Also like her blog, as well as several others, including Sohn's forecasting, which had this observation on the same subject of stimulus, but from a different perspective: tax reform and elimination of loopholes as stimulus:

"The U.S. tax code has too many loopholes, including the mortgage interest-rate deductions for homeowners and oil depletion allowances for the rich oil companies. Each loophole hurt economic growth because it goes against the market which allocates resources more efficiently. By closing the loophole, the productivity of the economy and economic growth improve. As more tax revenues are collected, the marginal tax rates can be lowered, boosting incentives to work harder for individuals and businesses. Again, productivity increases boosting economic growth and tax revenues. The budget deficit would shrink. Unfortunately, the power of special interest groups in the U.S. should not be underestimated. So, the probability is not high."

There's a simple reason people are lending to the U.S. government at those rates:

There aren't higher rates of return elsewhere!

And no, I don't trust the government to find them.

This question is framed very poorly. It is not a question of whether we should borrow and spend at negative rates, but of whether we should borrow and cut taxes at negative rates. That way, you avoid blowing money on whatever cockamamie scheme congress comes up with.

If real rates are -2%, you suspend payroll taxes and corporate income taxes until rates go back up and households and corporations fix their balance sheets. You could have implemented this policy after the crash and stipulated that it stay in place until real per capita GDP regained it's 2007 peak (it still hasn't).

Last time I looked we were already doing that. Recently watched Warren Buffet's interview with Charlie Rose about his tax increase proposal. One of the points he made was that the government spending 10% more of gdp then it was receiving in taxes and 5% more of gdp then the historical average of government spending WAS significant stimulus. If the suggestion is we re-allocate some of the current budget away from military/mandatory spending toward infrastructure, I would happily go along with you. Unfortunately, given the strength of interest groups opposing such changes, I suspect such a proposal would be DOA in Congress.

No, we aren't doing that at all. How much of the real increase in federal spending since 2008 has gone towards tax cuts? A small fraction. Adding up all real increases in spending since 2007, we have $2.2T in 2005 dollars. Of that, about 29 or 30% were tax cuts -- this includes the Bush Tax Cut Extensions.

Nick, in case you missed, Congress is meeting to cut $1.2 trillion in order to "save the country", so you proposal for tax cut deficits will get nowhere.

And, on top of it, state and local governments are laying people off, making unemployment higher, while our federal budget keeps cutting current spending, increasing fiscal drag.

Yes, if only government had more money to spend.

Every penny of government spending must first be extracted from profit realized by the private sector. Profits are down, so tax revenues are down. If the government wants to solve its funding problem, it needs to enact policies that increase private sector profits. Sucking in more private sector capital won't do the trick.

But, it can borrow at a price lower than the private sector. And, the private sector is neither borrowing nor investing, nor will they, unless they see demand from someplace.

In case you missed it, government contracts for most of its goods and services with businesses and the private sector.

"And, the private sector is neither borrowing nor investing, nor will they, unless they see demand from someplace."

And that demand is to be funded by negative rate loans, right? Does it occur to you that there's some reason the rates are negative? That maybe there is something seriously wrong out there that needs to be liquidated first?

Sure, the funding of public sector projects that would have to be done in the future anyway SHOULD be done with negative rate loans.

This is the weakest argument I have ever heard

Also, it seems very selfish and condescending.

"How many people would turn it down, even with their spouse in charge of spending the money?" I wouldn't take the deal, ESPECIALLY if my spouse was making the decisions...

TC's point doesn't seem to have anything to do with the rate of interest. The proper title of the post should have been:

"Should government spend more"

His argument only discusses his priors about the skill of government at spending money, whether taxed, borrowed or printed. He dresses it up with a fun thought experiment, but takes it no further.

The rate of interest matters only if that alone might influence decision-making. I've never heard that the concept of "hurdle rate" (the interest rate at which a particular spending decision begins to make sense) is ever used in government.

"and the non-crazy cousins screw up pretty often too." So because government screws up often enough, we should just abandon it. Wish they had thought about this before the Iraq war. Or various military boondoggles (Sgt York?). Prof. Cowen's argument seems to imply that the U.S. has lost it; after all France and Japan can have high-speed rail, and now China; admittedly with government subsidies. But they are better countries for it. Should the U.S. have undertaken to build the Interstate Highway System at the time? Undoubtedly, as with any big construction project, there were cost overruns somewhere (and I heard Henderson of the Hoover Institution argue this morning on KQED that the U.S. government should have stayed out of the interstate business, because it displaced all these toll roads private companies and the states were about to build, and anyway, Eisenhower was fixated on roads because of a road trip he had taken ~30 years before then. Seriously.).

In consideration of America's crumbling infrastructure, the proper thought experiment is not the one with crazy cousins, but that of a family with a house with a leaky roof. Should we fix the roof now, that we can borrow cheaply, and one of the kids gets a job as an helper with the roofing company, or should we wait until rot of the house frame sets in?

I forgot to add, that in the current political environment, the 'crazy cousins' are mainly Republicans, who want to give away money to, e.g. the drug industry (by preventing Medicare Part D to negotiate for the same volume rebate as the Veterans Administration gets), start optional wars, keep tax rebates for billionaires.... and as we hear from Mr. Cantor, offset spending on disaster cleanup by cutting programs benefiting non-millionaires.

Comments for this post are closed