In the Nation's Interest

It’s Krugman Time: ‘Time To Borrow’

In the past week, I have received an unprecedented number of CED member requests to review a newspaper opinion column. Those CED Members asked that I opine on Paul Krugman’s piece, “Time to Borrow” which appeared in the August 8 New York Times. Because your wish is my command (it says so in my job description), here goes.

Krugman has written several similar columns in the past, but this one is quite specific. The nation needs “sharply increased public investment in everything from energy to transportation to wastewater treatment.”

“How should we pay for this investment? We shouldn’t… Right now there is an overwhelming case for more government borrowing… The federal government can borrow at incredibly low interest rates: 10-year, inflation-protected bonds yielded just 0.09 percent on Friday… Spending now would mean a bigger economy later, which would mean more tax revenue.”

Sounds tempting. But before I sign on the dotted ballot, let me go over Krugman’s rationale for spending more on infrastructure without paying for it.

Item 1: Good old-fashioned Keynesian economics. “Spending now would mean a bigger economy later…” Eight years ago, sure. Right now, we economists would be testing our knowledge. We have not seen an economy like today’s in the living memory of, well, everyone who is now living. Unemployment this low, but with inflation this low, and with growth this sluggish and productivity growth in the dumps? What happens if we mash the fiscal accelerator? Forward motion, or more saving by the most well off, or purely inflationary wheelspin? Honestly, I don’t know. I’d love to see it, if we were playing for matchsticks. I have to chalk this point up as questionable.

Item 2: Infrastructure spending as Keynesian economics—timing. This is a worry that Krugman does not raise. Advocates of infrastructure spending always say that there are plenty of “shovel-ready” projects to do for quick-moving economic stimulus. Critics say that for any significant projects, choosing, planning, permitting, hiring, ordering materials, and so on take such a long time that the stimulus payoff of infrastructure spending is often too little and too late. But not exactly, because all of that delayed spending can finally hit the economy after it has already recovered—when the stimulus does more harm than good. Would that even be possible, in this languid economy? Good question. But it might suggest that if you really want to stimulate this economy—now, not at some unknown later date—you might choose some other way.

Item 3: Infrastructure spending as Keynesian stimulus—labor. Krugman didn’t mention this one either, but I hear it all the time as an argument for a “crash program” (ugh) in infrastructure investment. The argument goes that construction employment is way down from before the financial crisis, and so there are lots of construction workers standing idle. Putting those idle resources to work, in theory, is virtually free to the economy. We wouldn’t be taking those workers away from other jobs, so there would be no “opportunity cost” to putting them to work on infrastructure. Besides, they already have the skills, and so could step right into the new infrastructure jobs. Why not put them to work?

Well, the premise doesn’t quite stand up. It turns out that all those construction workers who lost their jobs weren’t in infrastructure-related construction (or “heavy construction” as it is called in the employment statistics), they were in home construction. Now, it may be that people who strung electrical cable and covered it with drywall would be decent candidates to drive steamrollers and pour concrete, but they aren’t prepared to step right into the job. Significantly expanding infrastructure spending might well take more time than is often assumed, and might involve bidding workers away from other jobs rather than instantaneously increasing employment.

Item 4: Infrastructure as stimulus—capital. Again, this issue is not mentioned in the Krugman piece, and in fact it is a negative for his case. And again, this one relates especially to a one-time “crash program.” Infrastructure construction takes a lot of heavy equipment. So consider what that means for a “crash program.” You have to buy a lot of new equipment to scale up your level of activity for a while; and then when the stimulus program is over, you park it. This would not be a very efficient use of resources.

Item 5: Productivity. “…Investing in infrastructure would clearly make us richer.” In this statement, Krugman probably includes the notion that more infrastructure would improve the productivity of our workforce. And to some extent, this is true. But there seems to be a widely held exaggerated notion of the magnitude of this payoff, at least in terms of its effect on the federal budget.

One opportunity rated highly by many pundits is investment in highways. With more capacity, people and goods would get to their destinations sooner. With repair of worn-out roadways, cars and trucks would incur less damage and need fewer repairs. With bridge repairs, people wouldn’t fall into rivers. It sounds like a no-brainer. Print those bonds, and let’s get going!

Well, this needs a little more thought. Is highway repair and construction a good idea, even necessary? There is plenty of supporting evidence. Engineering studies certainly agree. But Krugman’s proposition adds one more layer: that we borrow a lot of money to start a crash program. It will pay for itself, if we need to worry about the debt at all (more on that later). That add-on proposition is a big, big stretch.

How about added capacity? A highway build that saves each affected commuter 40 hours a year would be a romping, stomping success. It sounds like a free vacation! But for a typical 48-weeks-per-year worker, that is five minutes per commuting trip. That amounts to a little more lingering over the morning coffee. And OK, a little less stress, too. Would it be a good idea, and should we do it? Depending on the cost, quite possibly so. But paying for itself? In the federal budget? Forget it.

If a bridge collapsed, it would be a tragedy. Should we repair our deteriorating bridges? Absolutely. But would reducing the not-quite-low-enough probability that a bridge will fail increase our nation’s productivity? And pay for itself in the federal budget? Not at all.

In short, and whimsically, we would have to prevent a lot of front-end-alignment jobs to pay for a multi-billion-dollar infrastructure stimulus program. But if the highway investment actually did obviate the need for so much repair, of course, the benefits would be received by individuals, not by the federal government. Tax revenue would go down, not up, making it even less likely that the infrastructure program would pay for itself.

And we have already tried this approach once. The 2009 stimulus bill, of its $787 billion of additional deficits (through both spending and tax cuts), added $308 billion in the form of discretionary (appropriated) spending—the category of spending that would comprise a Krugman-style bill today. If such an increment to public investment would have increased worker productivity, we would be swimming in productivity growth. But we are having this conversation now precisely because productivity growth is so slow, and worker income growth is slow along with it.

So far, we have been talking about Krugman’s case for why his proposed public investment bill would be efficacious. Now, let’s think about his case for why the addition of the cost to the nation’s bar tab would not be harmful.

Item 6: Low interest rates. “…The federal government can borrow at incredibly low interest rates: 10-year, inflation-protected bonds yielded just 0.09 percent on Friday.” That interest rate is indeed very low, virtually zero, and 10 years to a condemned man is almost forever. Why not borrow the money? Well, for starters, Krugman pitches a bond that is the equivalent of an adjustable rate, not a fixed rate. Inflation right now is very low, quite possibly below the Federal Reserve’s target zone, precisely because economic growth is historically slow. Ironically, if Krugman’s program works, it will undermine itself in this respect. If economic growth accelerates, inflation will rise, too, and the cost of servicing these bonds will go up. This is the “Catch-22” we find in the fiscal corner into which we have painted ourselves. (Conventional, fixed-rate Treasury bonds are still at historically low rates, but well above the inflation-protected “TIPS.” The 10- and 30-year Treasury yields last Thursday were 1.57 and 2.28 percent, respectively.)

Item 7: We don’t already have too much debt. After conceding that our debt has accumulated to a big, imposing dollar number, Krugman argues that “…everything about the U.S. economy is huge, and what matters is the comparison between the cost of servicing our debt and our ability to pay. And federal interest payments are only 1.3 percent of GDP, low by historical standards.” Here, Krugman is back in the same Catch-22 as he was one paragraph ago. For all practical purposes, almost the entire public debt is on variable rates. As the following chart shows, almost one third of the debt will mature within one year. About another one third will mature between one year and five years hence. Once economic growth gets off the floor, whether because of Krugman’s program or not, interest rates—and the cost of servicing the debt—will increase by orders of magnitude. The massive debt that we already have accumulated gives those interest rates enormous leverage over the federal budget.

Item 8: It doesn’t matter if interest rates rise. “But we’re talking about long-term borrowing that locks in today’s low rates. If 10 years isn’t long enough for you, how about 30-year, inflation-protected bonds? They’re only yielding 0.64 percent.” Krugman intones (on paper, which is a neat trick) “inflation-protected” as though it were protecting the borrower. It isn’t. It protects the lender (that is, the bond buyer). If Krugman’s program works, and growth and inflation rise, the cost of servicing those bonds will rise, too.

But there is more to it than that. Just how big of an infrastructure program does Krugman want? A cost equal to only a small percentage of the GDP would not have the kind of impact he seeks. Does he want something like the $308 billion discretionary-spending component of the 2009 stimulus bill? (And that was only a little more than one third of the total size of the 2009 stimulus, and would be only about 1.7 percent of GDP today—not huge by macroeconomic policy standards.) If so, he might want to consider that the total outstanding stock of Treasury bonds with greater than 20 years’ maturity is only about $900 billion. Stuffing a $300 billion public investment program into long-term bonds would increase the total amount of these bonds on the market by one third—a massive change over a short period of time. In January of this year, Treasury held one auction of 30-year bonds. It raised $13 billion. Krugman’s auction would be almost 25 times larger. In February, Treasury held one auction of 30-year inflation-protected bonds, to which Krugman refers specifically and approvingly. In this case, the auction raised only $7 billion. So if Krugman wanted that bond’s really low stated yield (which for the Treasury would be subject to inflation risk), his auction would be 44 times as large.

Infrastructure stimulus fans seem to think that the Treasury can sell any duration of securities that they want, and in any amount, and the market participants will just gobble them up. In the kinds of volumes that Krugman is talking about, they won’t. Even spread over a longer period of time, these kinds of numbers, restricted to the very long end of the maturity curve, would be highly disruptive. So the public debt will remain of relatively shorter duration, and therefore essentially at adjustable rates. And when the economy recovers, debt service will become, as forecast, the fastest-growing spending line in the budget.

And if Krugman wants to say that he can add the cost of his program exclusively to the long end of the maturity structure without disrupting the financial markets because the money will flow much more slowly, he is simply admitting that the economic impact of his program will be much slower and more diluted than he had claimed. He is doggoned if he does and doggoned if he doesn’t on this question.

So where are we on infrastructure investment?

Let’s face it: Our policymaking has been short-run centric for much of the last three-and-a-half decades. We have been managing for cash, and when you manage for cash, you defer investment and you defer maintenance. We need to right the ship (and tar the hull). On this, Krugman is, in my judgment, absolutely correct.

Where he goes wrong, in my view, is in taking the attitude that public investment is so right that it must be free. Like the supply-side tax cutters who tell us that we must somehow endure the pain of huge tax cuts, so the supply-side spenders say that if we will only suffer through an enormous increase in public spending, it will all turn out for the best in the end. Self-indulgence pays.

Would that it were true.

And yes, we have painted ourselves into a fiscal corner. Catch-22 is upon us. We cannot risk larding on the debt for spending that might help the economy. Taking on the massive debt that we already have has restricted our options.

We should create a solid, fiscally responsible infrastructure program. The construction industry should be fully employed, and expanded according to a manageable schedule such that the taxpayer gets full value as projects are identified, vetted, prioritized, planned and executed. But as far as the financing is concerned, let’s be realistic, rather than staking our collective futures one more time on the arrival of the tooth fairy.