In the Nation's Interest

Things You Learn When You Work on the Public Debt

You certainly have noticed in the last week or two that the “national debt” – in precise technical language, the debt held by the public – has become a topic of considerable interest (pardon the pun) in the election news.  There is a fair amount of loose talk on the subject.  Given that the debt is now $13.8 trillion – and change – it might be worth a brief conversation.1

A word of personal background:  I spent a fair amount of time hanging out with the public servants in what used to be called the Bureau of the Public Debt (now a part of the Bureau of the Fiscal Service) within the Department of the Treasury in the late 1990s, when we were actually paying down the nation’s debt.  That required active collaboration between the Treasury and the Office of Management and Budget (my employer) – two agencies whose staffs in the 1989-1992 years were under standing orders not to speak to one another.  I have no idea what that was like, but I am confident that cooperation was much more fun.  I met some great people, and I learned a lot.

And the single core fact of what I learned in those days, on point with the loose chatter that you are hearing now, is that the public debt, and its role in the financial marketplace, are quite different from the nature and role of the debt of any private entity.  (For a very broad answer to the oft-asked question, “Why can’t you bozos run the federal government like a business?”, go to https://www.gpo.gov/fdsys/pkg/BUDGET-2000-PER/pdf/BUDGET-2000-PER-4-2.pdf and read the box that begins on the second page of the document – the printed page numbered 18.)  Assuming that the federal government could manage its debt using the instruments and the methods that might be profitable in the private sector is, well, misguided.  How is that true?  Let me count the ways:

1. Unlike many private entities (and many state and local governments), the federal government follows unitary financing almost exclusively.  That is to say, apart from the maturity durations, all Treasury securities are identical and are fully interchangeable.  The Treasury follows that long-standing practice because it saves money for the taxpayer, and it smooths the functioning of the financial market.  If you buy a Treasury bond, you know that it will be highly liquid.  If you unexpectedly need to sell it, you will not be stuck with a white elephant with some unique covenants or restrictions.  You can take your bond to a marketplace where many identical Treasury securities are bought and sold every day, and so you can get a bid at a fair, competitive price.  That liquidity means that you will be willing to bid more for a Treasury, all else equal, which means that the taxpayers get a better price when they (implicitly) sell their bonds to borrow money to finance their government.  Meanwhile, the cost of administering the initial sale (auction) of those bonds is less, because they are all the same, and therefore there are economies of scale for the government – which is to say, the taxpayer.

This is as distinguished from project financing, in which an entity borrows money through a security that is specifically tied to the purchase of a particular money-making asset (or group of like assets).  A state might sell highway bonds that are secured by the tolls expected to be paid on that highway.  If the traffic load is lighter than anticipated, the bondholders take it in the neck; they do not have a legal right to a wing of the state capitol to make them whole from their losses.  Likewise, a real estate developer (to choose a line of business at random) might mortgage a building through a loan that is securitized only by the building in question.  If the building proves to be a poor investment, then the lenders and the developer must strike the best deal they can to remedy the revenue shortfall.  The lenders by the assumption in this hypothetical situation have no claim on the developer’s other property or wealth, and must negotiate for whatever partial reimbursement they can get.  If they would not be willing operators of the property, and they would take a capital loss if they assumed ownership of the property and then tried to sell it, they might be over a barrel in a negotiation with the developer.  In the worst of all possible worlds, the developer might even enter into the transaction anticipating that he or she might be able to exploit the lenders in a likely or possible renegotiation over the debt.

The implication for the current discussion is that the entire public debt, unlike some private debt, is associated with the entire government as an enterprise.  Money is fungible, and the (unitary) debt is fungible.  It cannot be said that any particular bond financed some particular aircraft carrier, a particular mile of the interstate highway system, or one month’s worth of food stamp benefits.  And as such, the interest obligation on that debt must be met just to keep the lights on.  The federal government cannot strike selective deals to scratch savings on parts of its outstanding debt, in return for bargaining about particular federal assets – what some might see as a potential “out” on the debt burden that might have been hinted at in some of the recent chatter.

2. And that unitary federal debt is really, really big.  The United States is the world’s largest borrower by a long, long way.  This was true throughout our lifetimes, but it is especially so over this century thus far.

The very size of the public debt should instill great caution in policymakers.  Any miscalculation could have enormous consequences.

3. That enormous US public debt plays a unique role within our nation and around the world.  For decades, Treasury securities have been the global standard for safety and liquidity.  As such, they make the ideal reserve asset, and they have been so employed – by private financial institutions, governments, and numerous other entities and institutions.

Consider what that means.  Any misstep by the federal government that called the value of Treasury securities into question would cause a ripple – check that, a tsunami – throughout the entire world.  Financial institutions – insurance companies, pension funds, banks, brokerage firms – that hold high credit ratings precisely because they have sufficient reserves in hitherto rock-solid US Treasury securities could be downgraded.  Institutions that rely on them could be downgraded.  And so on.  Missed payments because of that turmoil would build and extend the wave.  All of these institutions would be trying to raise cash and accumulate high-quality assets to restore their standing – all at the same time.  This is what we call a financial panic.  The dislocation of 2008 could look like a mere sneeze by contrast.

This fear should arise when we hear people talking about the option to “renegotiate” the debt.  Taken literally, this would mean asking individual bondholders to accept less than 100 cents on the dollar.  If this is what people mean (or meant), they should realize that the consequences of even opening the door to this possibility could be catastrophic.  More recent discussion seems to have moved from “renegotiation” to “refinancing.”  See (4) and (6) below.

4. So to avoid such a dislocation, some today talk about saving money on debt service to defuse the threat.  And in truth, minimizing the taxpayers’ cost of financing the debt is an important policy objective of our debt management.  But it is not the only objective.  Given the size of our outstanding debt, and the status of our securities as the “riskless” standard of the world, we manage our outstanding debt so as to stabilize the global financial markets, in turn to facilitate and even stimulate commerce and trade domestically and internationally.

One of the tools suggested to cut the cost of debt service is to lengthen the debt’s average maturity, by selling fewer short-term bills and notes, and selling more long-term bonds.  And in fact, since the financial crisis began, the Treasury has lengthened the maturity of the outstanding debt fairly significantly.  But there are limits.

First of all, part of the Treasury’s objective to keep the markets sound and stable is to define a yield curve for the “riskless” asset all along the maturity structure – from three-month bills to 30-year bonds.  OK, you say, drop that objective.  Move to 30s today and lock in those current low interest rates for a few decades.  Leave the financial markets to fend for themselves.  What do we care?

Well, we should care; the soundness of the financial markets matters.  But beyond that, the Treasury cannot sell what the market does not want to buy.  Yes, insurance companies and pension funds want 30-year bonds, to match their assets with their future liabilities.  But the markets have demanded huge volumes of three-month bills over the last eight years, even as monetary policy has held that rate right down virtually at zero.  Times are uncertain, and many people want the certainty that they can redeploy their capital without capital loss at an early date.  Tell those people that you have a Treasury security for them, but that the maturity is 30 years instead of 90 days, and they might buy it – but at a price in terms of yield.  Don’t expect today’s 30-year interest rates to prevail in such an altered market.  And keep in mind that our debt is so large in total that shifting the maturity structure significantly will entail selling a lot of paper at those longer maturities.

And don’t forget that in trying to lengthen Treasury maturities, you will be placing a bet.  As this is written, the annualized market yield on a three-month Treasury is 0.22 percent; the yield on a 30-year bond is 2.71 percent.  Withholding one three-month bill and instead auctioning one 30-year bond increases the debt service on that security right now by a factor of more than 12 (assuming that shifting the composition of supply along the maturity structure has no effect on yields).  How confident are you that the economy will strengthen enough in the near term to raise short-term rates above 2.71 percent annualized, and by when?  So lengthening maturities may or may not actually prove to be a money-saving bet within the foreseeable future.

(On a separate but related topic, this is one more reason why those who say that “with today’s low interest rates, this is the perfect time to borrow money and invest in infrastructure” are not necessarily on the mark.  First, we already have so much debt that incurring any more debt at all may be unwise.  But second, we cannot finance all of that infrastructure at today’s low 30-year rates; the market will demand that some of that additional borrowing be undertaken at three-month rates that buy us no financial security whatsoever.)

5. There is one more point of misunderstanding in the common assessment of the relationship between our debt-service cost and our economy.  Some seem to believe that our debt situation will improve substantially if the economy grows more rapidly, and that our concern should be if the economy performs poorly.  The common language seems to be that we should engage in extraordinary action – perhaps “renegotiate” the debt – if the economy “tanks.”  Well…  This is generally true.  But it is not as clear-cut as it once was.  And this particular change in our circumstances is troublingly indicative of the box into which we have stuffed ourselves.

The relationship between our debt service cost and the state of our economy is an amalgam of two contrary forces.  As the economy strengthens, revenues tend to increase, and income-sensitive budget costs (quintessentially, unemployment compensation benefits) tend to decrease, thus reducing the budget deficit, the increment to our debt, and therefore future debt-service costs.

But on the other hand, as the economy strengthens, the demand for credit tends to increase, which drives up interest rates.  Meanwhile, faster growth tends to put upward pressure on prices, which might raise the ire of the Federal Reserve.  Therefore, again, interest rates tend to rise.  That increases the cost of servicing the debt that we already have accumulated.  The net effect of faster economic growth on net interest costs, therefore, is the balance between these two influences.

If we had zero accumulated debt, the balance between those two forces would be unambiguous.  Higher interest rates applied to our zero debt would cost us nothing.  Our financial standing would only improve as the annual deficit shrank (or the annual surplus increased) from the higher revenues and lower outlays.

But as of now, relative to just about any time in post-World War II America, our outstanding volume of debt is large.  Therefore, as the economy is perceived to improve, the cost of servicing the pre-existing debt is larger relative to the benefit to the current non-interest budget.  Mind you, that balance likely is still favorable, and no budget wonk should walk around praying for slow growth to keep interest rates low.  But it will take more growth today to get the same amount of end-of-the-pipe budget improvement than was the case in the pre-financial-crisis years.

So we should not assume that faster economic growth ensures an escape from debt troubles.  If the debt is large enough, and interest rates are high enough, even a growing economy may be outpaced by the rate of growth of the debt – in other words, the debt-to-GDP ratio can grow even if the GDP is growing.  And if we find ourselves in that space, economic growth will not go on very long.

Again, this is a warning of the vulnerable position into which we have put ourselves.  We are in a bad place.  And if we ever again in our lifetimes find ourselves with options, we should choose those that keep us far away from where we are today.

6. Another misconception:  Some seem to think that the federal government could use a financial tool available to homeowner and business borrowers and “refinance” the debt.  As a simple example, most home mortgage loans can be pre-paid without penalty.  So if you have an outstanding 6 percent mortgage, and the current going rate is 3 percent, you can pay off your 6 percent mortgage by taking out a new mortgage at 3 percent and cut your debt-service costs in half.

The federal government does not have that opportunity.  There is no provision of bond language that gives the government the option of pre-paying bonds at par.  If the federal government wants to “refinance” its outstanding debt, it must sell bonds at current market rates to purchase outstanding bonds at current market rates.  In other words, every such transaction would be a wash – minus administrative and transactions costs, meaning that such an exercise would be a money-loser for the Treasury.  (You will realize quickly that this simple analysis does not apply to the Treasury’s “reverse auctions” in the late 1990s, actually buying back debt only.  This is because at that time the federal budget was in surplus, the Treasury had excess cash, and so there was only one transaction to retire debt without a precisely offsetting sale of a bond to raise the money in the first place.  Thus, even though some chose not to undertake the analysis at the time, even a purchase of bonds trading at premiums because of then-low interest rates saved the taxpayers money.)

So some might ask, why not put covenant language into all future Treasury bonds to allow the federal government to repay those loans without penalty?  The Treasury certainly could.  But investors would not be willing to pay as much for those bonds – or in other words, their interest yields would be higher.  It might even make some financial institutions significantly less willing to hold Treasury securities as reserves or collateral, because the risk of unanticipated repayment would make those institutions’ operations and planning processes more complex.

7. One way to wish away our current public-debt concerns is to fall back on the government’s power of money creation.  The federal government cannot go bankrupt, because we can simply print the money to meet our debt-service and debt-redemption obligations.

True enough.  But meaningless.  To borrow a recently used figure of speech, this choice is like that between a bullet and poison as the cause of one’s death.  The economic consequences of government bankruptcy would be hard to distinguish from those of hyperinflation driven by an infinite rate of money creation.  (And though I am not expert on this subject, those who are far more so than me say that the short maturity duration of our outstanding debt means that government likely cannot print money fast enough to stay meaningfully ahead of bankruptcy strictly defined.)

8. A final note:  The one topic that to my surprise has not been mentioned in this public chatter is the implication for Social Security.  Social Security’s most avid defenders – those who believe that promised benefits must not be cut, at least until the split second when the program’s trust fund is fully exhausted – should be seriously troubled by any mention of “renegotiating” the public debt.  To the literalists on the subject, the trust fund is iron-clad because it holds Treasury securities that are “backed by the full faith and credit of the United States.”  Well, so are the marketable Treasury bonds that the current chatter suggests will be “renegotiated.”  If the value of those securities is negotiable, why not the Treasury “special certificates” in the trust fund?  This entire conversation should bring home to the avid defenders of Social Security that the “full faith and credit of the United States” will be of debatable value if the security of the general fund – not just the trust fund – comes into question.

So what have we learned?  Experts in government finance should respect experts in business finance.  And vice versa.  The skills of each group may be transferrable to the other – but the fact base is not the same.  There is always room for new ideas, but there can be valid reasons why the customary practices differ between the two disciplines.  And most specifically, we should not assume that there is some easy way out of the current public debt morass.  The stakes are high, and the consequences of ill-considered experimentation could be more serious than the conventional wisdom may realize.


1. Footnote, and example of loose talk:  You may be puzzled about that $13.8 trillion number.  Didn’t you just hear someone else talking about $19.1 trillion – and change?  Well, the Treasury owes $13.8 trillion to the public (including the Federal Reserve).  The other $5.3 trillion of the “gross debt” or “debt subject to limit” (subtly different from one another) is what the federal government owes to itself – including its Social Security Trust Fund, the Medicare Trust Fund, the Highway Trust Fund, the Abandoned Mine Reclamation Fund, and all of the 100+ other happy members of the extended Fund family.  Given that the recent conversation is about debt-service cost and default, and all it takes for the federal government to be current to itself is to change an internal bookkeeping entry, I don’t quite know why anyone would make reference to the $19.1 trillion.  Except that the bigger number might be scarier.  Though personally, I can’t afford either one.