US Faces Low Borrowing Costs—Not Free Lunch

August 14th, 2012 at 6:00 pm

Those of us who want to see more temporary government spending on jobs programs right now (I’d call it “stimulus” but that’s a dog whistle these days) often cite the very low interest rates of government bonds as yet another good reason—along with the high unemployment rate—to do this.  Last month, the interest rate on a 10-year government bond averaged 1.5%, the lowest level on record going back to the early 1950s, so the cost to the government of borrowing right now is very low.

But as I’ve pointed out before, the reality of this is more complicated.  I stumbled on some data from the Office of Management and Budget the other day that helps make the point (h/t: RK).  It is still the case that the government should be taking advantage of low borrowing rates but the advantage may be less than many who write about this think it is.

The reason is because the government doesn’t pay off such debts after a year or two.  It rolls them over and thus, under the plausible assumption that interest rates go up in the future, it pays that debt off at higher rates down the road.

The figure below presents two ten year trends in what the Treasury expected to have to pay in interest rates looking out from 2002 and 2013 (these are weighted average rates representing Treasury’s overall mix of financing, with maturities ranging from 3 months to 30 years).  The key point is that the administration assumes—as does CBO—that the average interest rates faced by Treasury are expected to rise from less than 1% now to about 4% by the end of the 10-year budget window.

Source: Calucations based on OMB data.

Of course, these forecasts could be wrong, as you can see.  Back in the early 2000s, OMB thought Treasury would pay a rate of over 5% in 2011, as opposed to the far lower rate that actually prevailed that year.

That said, how would the stream of debt payments for a stimulus package differ under these two rate regimes?

Imagine Congress came to its senses and passed $300 billion in stimulus spending, $200 billion next year followed by $100 billion in 2014.  According to OMB rate estimates shown in the blue line above, interest payments would amount to about $100 billion over ten years.  Applying the other set of interest rates from the figure above—think of those as the set that would prevail in less unusual economic times—interest costs would be about $150 billion.

So, the savings from taking advantage of today’s low rates would be $50 billion.  That’s by no means chump change, but it ain’t a free lunch either.

Print Friendly, PDF & Email

12 comments in reply to "US Faces Low Borrowing Costs—Not Free Lunch"

  1. jim says:

    All the more reason to have the Fed buy the debt. Doesn’t that essentially solve the debt problem?

  2. foosion says:

    The 5 year TIPS is currently yielding -1.15% real (after inflation). This produces genuinely free money – investors are willing to pay the government 1.15% real per year to hold their money for 5 years.

    The larger point is that everything is on sale now. There is necessary maintenance, infrastructure spending, etc. that has to be done at some point. Do we do it today, when borrowing costs are low, there are a substantial number of unemployed who could do something productive (and would do it at a low price) and materials are relatively inexpensive, or do we do it later when everything is more expensive? Fixing stuff isn’t free, but is cheaper than usually. It has many financial and non-financial benefits (e.g., stimulating the economy, giving people something productive to do).

    We’ll be better able to pay for stuff and to repay the debt when the economy is stronger than today. Yet another benefit from borrowing today.

  3. rjs says:

    so, youre making a good arguement for issuing more long term bonds…

    if we invest in infrastructure or education that would have more than a 3% long term return to the economy, any borrowing we do turns a profit…

    • Chris G says:

      >so, youre making a good arguement for issuing more long term bonds…

      This seems exactly right to me. Issue long bonds at a low nominal interest rate and then implement expansionary fiscal policies which, if effective, will have the effect of driving the real rate down. Voila. Free money. (And if something seems too good to be true it almost certainly is…)

      With that in mind, if the Gov’t finances stimulus efforts with short term debt they need to turn over then it seems to me you lose the ‘benefit’ of current conditions. But isn’t that exactly what Operation Twist did? (Sorry, I should be paying closer attention.) I’m still a bit puzzled about how, in practice, Operation Twist was/is supposed to goose the economy.

  4. Frank says:

    I agree with the post. I also agree with the MMT-ers on some things but I still don’t get how they deal with a scenario of rising interest rates pushing up debt service on national debt that they say does not matter. I think they say that the Fed can control interest rates so this is no problem but what if the Fed needs to have interest rates rise to control inflation at some point.

    • RayW says:

      The debt does not turn out to be an issue because the spending is not dependent upon the borrowing. Congress sets a spending amount and the government spends it. It is an inconvenience/insidious plot (smile)/contrivance/etc that the government sector must offset spending with revenue.

      The U.S. government can spend on available real resources and pay interest on bonds denominated in U.S dollars without constraint. The issue is that once the U.S. government begins to compete with private sector desires to spend on real resources the policy considerations apply. We are not even close to scarce competition for real resources. Unemployment is high, we have extensive unmet infrastructure, medical service, and educational needs, and our production capabilities are largely idle. I think we do have some agricultural surplus, but that is due to how we choose to distribute it. The world has very many people that could eat what we let mold.

      This is observable information. I challenge anyone to chart the mechanics of our government’s spending in any other way. The press is dishonored by not challenging the debt crisis “wisdom” and asking for facts rather than notions and feelings.

      The “looming debt crisis” is a notional projection based upon naivety, foolishness, or malicious duplicity. Pick one.


  5. Paul Papanek says:

    Professor – Not sure I see why our long-term costs would necessarily have to rise even to 4%, when the yield on 10 year treasuries is 1.82% and even the 30-year is now at 2.92%.

    (Link –

    No need to roll those over, unless we feared that a stimulus that finally was adequate to the job of plugging our gap in aggregate demand couldn’t get us back to full employment even in 10 years.

    Let’s imagine that America could implement a new two-trillion dollar stimulus over the next two years, using only money from new treasuries. Using a recurring multiplier of 1.5, and federal tax rate of 15 to 20% on any GDP increase, the Treasury would break even, and repay the principal itself, within 5 years. The accumulated interest on the Treasuries by then would be less than $200 billion, with projected increased annual tax revenues of at least $50 billion to $100 billion starting after year 2, and growing. Even with a more pessimistic multiplier of 1.3 (and factoring in some drag, and not even counting the savings from lower safety-net payouts), we’d get to break-even at 9 or 10 years.

    With a more sensible overall federal tax rate of 24% of GDP, targeted progressively at idle money concentrated among the wealthy, we’d get there even quicker.

    Yes? No? (Thanks for the terrific ongoing posts.)

    • Jared Bernstein says:

      First, two trillion in stimulus–13% of GDP!–is way outside the realm of the possible even with a willing Congress. But your larger point is a good one and your math is in the spirit of a recent paper by Summers and DeLong who made a similar argument.

      But the problem is that Treasury always rolls over debt–which is now in the $11 trillion neighborhood. So any spending that adds to that will generate interest costs as per the prevailing rates that prevail within the budget estimation period–say, 10 years.

  6. JimZ says:

    If the country has the good fortune to have rates rise to the 4-5% in a few years, that will mean that the economy will have rebounded with strong demand, lower unemployment, etc., presumably as a result of said stimulus spending financed by the new borrowing. Win-Win!

  7. David says:

    The bond market sure doesn’t expect rates to approach 4% in the next five years. Otherwise, the interest on 30 year bonds would be far higher than 2.9%.

    • Jared Bernstein says:

      Good point.

    • Chris G says:

      So who’s buying 30 year bonds? Seems to me – a layman – that there’s substantial risk in buying a 30 year bond at 2.9 pct. Thirty years is a long long time. If things return to what was normal then you could lose a significant chunk of change. Seems like buyers are accepting a heck of a lot of downside risk.

      Ten years at 1.65 pct amazes me as well from an acceptance of long term risk standpoint. If you’re just concerned about preservation of capital then why not go in for two years or three? Then you wouldn’t be handcuffed if things get better. [suppresses laughter] Okay, so things aren’t likely to be much better in 2-3 years but how about five?