The national sub-prime borrower

Here's a bemusing statistic to follow on from the earlier post.

Question: Given that the US federal deficit is expected to be $1.8 trillion this year, taking the total national debt to $12.8 trillion, by how much does the US government expect their interest payments on that debt to increase?

Answer: They don't, they expect interest payments to fall by 44%, from $253 billion in 2008, to $143 billion in 2009. And next year, when they expect to run a deficit of another $1.3 trillion, taking national debt to $14.4 trillion, they expect interest payments to fall further to $136 billion.

Confused? I am.

A. Interest rates

One explanation is relatively simple. Short-term interest rates have fallen. Tim Worstall recently provided a useful explanation of the yield curve. If they are able to borrow both the new debt and the rolled-over debt at lower rates than before, then interest payments might fall even if the principle has increased. Depends on the amount of the increase in the principle and the fall in the rates.

The average maturity of Treasury bonds is said to be around 5 years. Longer-term rates aren't that much lower now than they have been over the past decade. Yields on 5-year bonds are down to around 2.5% (from 5% a couple of years ago and 3.5% a couple of years before that), and get lower the shorter the bond, down to a fraction of one percent for 1-month and 3-month bonds.

If the Treasury borrowed much of the new and rolled-over debt on medium- to long-term bonds, net interest payments would be going up (or at least not down by anywhere near as much as assumed), as the increase in the principle would exceed the reduction in the rates on that portion of the existing debt that was rolled over. To achieve the reduction in interest payments that they assume, a substantial proportion of the existing debts must be coming to maturity this year and next, and must be renewed (along with the new debt) as short-term borrowing. So a large proportion of their debt will end up on short-term rates.

The trouble is, short-term borrowing needs to be renewed frequently, and short-term rates don't stay low in a recovery (nor do long-term rates come down). There is no scenario that I can think of where the economy rebounds to the extent that the Obama administration is assuming, from the point of view of the tax-revenue it is able to yield (see previous post), and short-term rates stay at a fraction of one percent. At that point, the recovery kills itself, because the interest that has to be paid on the projected national debt of $14.4 trillion in 2010 or $15.7 trillion in 2011 should be at least 3% and probably higher, given the strength of the assumed recovery implied by the tax revenue. That's $430 billion of interest payments in 2010 (compared to their assumption of $136 billion) and $628 billion if average rates are nearer 4% in 2011 (compared to their assumption of $254 billion).

Those discrepancies would mean either increased taxes to cover the extra payments, or increased borrowing to cover the increased deficit. The former would stunt growth just when it was starting to recover, reducing tax revenue in a vicious circle. The latter would further increase the cost of borrowing, making it necessary to borrow more, in another vicious circle. Taken at face value, it's hard to see how this house of cards can't collapse. It's Catch-22.

Ironically, it's a similar predicament to the situation that sub-prime borrowers got themselves into. The government is being tempted by the equivalent of teaser rates to borrow greatly in excess of what its assets are worth and what it can afford to pay back. Those rates aren't fixed for long, and are likely to be significantly higher when they come to renew (or their income significantly lower, if the economy hasn't rebounded in the way they project). And by then, they'll have racked up even more debts.

Unfortunately for the American people, there is no "jingle mail" option when its government can't afford the payments and its debts exceed the value of the economy. Or should that be "unfortunately for the bond holders", because the government will have to take one of the two options that are closest to "jingle mail": default or inflate?

B. Net interest

But this is where another explanation for the apparently improbable figures comes in. It's the explanation that has me really confused. The figure that the US government has to pay is not the whole of the interest payments due, but the net interest. The net interest takes account of the fact that the largest holder of government debt is... the government! It owes the money to itself. As this is effectively a transfer of money from one part of the budget to another, it nets-off these amounts and only counts the net interest (i.e. the interest payable to all external holders of government debt).

I can just about wrap my head around the origins of this circular notion. The US government first started to hold substantial amounts of its own debt after 1983, when the Reagan reforms to the Social Security budget left it with a surplus of social-security tax receipts over social-security outlays, and facing substantial future obligations. They used the surplus to buy Treasury bonds, which in a sense you could see as a sensible investment to provide for future obligations.

But it doesn't really stack up. Unlike corporate bonds, the money received by the government in exchange for its bonds is not invested in profitable assets that will provide the return to repay the bonds over future years. Whether used for current expenditure or capital investment, the government gets no yield from the expenditure, other than the hope that its measures will cause the economy to grow more quickly than it otherwise would have, which will allow for higher tax revenues. Ultimately, government bonds will have to be paid out of future tax receipts. And if the government hadn't borrowed from itself, the money for social-security obligations would have had to come from... future tax receipts. All that has really happened is that a convenient excuse has been found to make a surplus in one part of the government's activities available for current expenditure anywhere the government likes in its budget, and a slightly greater moral obligation has been placed on the government to raise tax in future to pay for obligations it was almost certainly going to pay for anyway.

And there was no real surplus. The US government's deficits ran at around $200 billion a year on average between 1983 and 1995 (in nominal terms, nearer to $300 billion in today's money), and at similar levels (in real terms) from 2002. Total government outlays were $800 billion at the start of that period, which should give some idea of the relative size of the deficit. In practice, this device was a way to disguise imprudent and spendthrift expenditure as somehow prudent and parsimonious.

The net effect has been to swell the US government's share of gross national debt from 17% in 1985 to 44% in 2007, while that gross debt has been increasing significantly (from $1.82 to $8.95 trillion in nominal terms, from $3.12 to $8.95 trillion in real terms, at 2007 dollars, or from 43.8% to 65.6% of GDP).

We have become familiar with the concept of Quantitative Easing since the onset of the Credit Crunch. It might be thought that this is a sign that the policy had been running longer than we realized. But for the purposes of calculating the net interest, the Federal Reserve is treated as external to government. Payments due to the Fed are included in the net interest figures. QE as we understand it today would have resulted in an increase in the net interest payments. The Federal Reserve was not until recently a big buyer of government debt - it held only 5% of outstanding securities in June 2008 (around 10% of that big blue chunk in the pie chart below). The government had not got its money through the relatively transparent (if misguided) process of Federal Reserve balance-sheet expansion. It seems that the federal government simply issued itself IOUs in order to increase the amount of money it appeared to have available to spend. Rather than spending directly the tax it raised, it used some of the tax to buy its own debt. The money was made available for expenditure in the same way as if it had spent the tax directly, but the government had made obligations to itself to pay itself money in the future (money that will have to be raised from taxation and spent on the budgets for which the debt is earmarked, in much the same way as if the debt didn't exist).

It seems a strange thing to do. But in any case, the government doesn't have the tax receipts to buy as much of its own debt as usual, so the Fed is indeed buying a large chunk of federal debt this year (i.e. the government will have to expend future taxpayers' money to repay money borrowed from the Fed, which the Fed created out of thin air, and which was used to allow the government to continue to run unsustainable deficits and indirectly to blow bubbles in the stockmarket). The government's share of its own debt is projected to increase by "only" $150 billion (to $4.3 trillion), whereas debt held by the public (including the Fed) is expected to increase by $2.7 trillion. Last year, debt held by the public increased by twice as much as in the previous highest year ($768 billion, compared to $382 billion in 2004). This year's increase is 3½ times higher again, taking the total debt held by the public from $5.8 trillion to $8.3 trillion. This means that the share of the debt that is paid out of taxes to recipients outside the government has increased by a significantly higher proportion than the increase in the gross debt. And it is expected to increase by a further $1.35 trillion next year and nearly $1 trillion the year after that.

This is the context against which the net interest payable (the interest payable on the debt held by the public including the Fed) is supposed to fall dramatically and remain relatively low. The principle amount on which net interest is payable will more than double between 2007 and 2011, the economy will recover so strongly that it yields record tax revenues, and yet interest payments will remain modest, apparently. I'll have some of whatever the Obama administration is drinking.

But someone around government does seem to have taken account of this in at least one way. The Social Security fund was already facing an underfunded future. For many years now, the trustees of the fund have been forecasting that their surplus would turn to deficit in 2017, and all bonds would have been exhausted by 2041, at which point there would be insufficient funds to meet the Social-Security obligations. This year, they have shortened these projections, so the fund is projected to be in deficit from 2016 and to be short of funds by 2037. This presumably reflects the lower rates of interest being offered on the debt it is rolling over. And those lower rates of interest are the results of the artificially-low yields being generated by the policy of QE. So QE is not just blowing a new asset-price bubble, destroying savers' values, failing to reduce costs for borrowers and to inject much cash into the real economy, it is also exacerbating the near-term deficits of the US government, and bringing forward the day that the Social Security fund will not have the funds to pay its obligations.

Interesting that the markets are doing so well, because it looks to me like the government's projections are obvious fantasy, and the US economy cannot remotely afford the stimuli that the markets are relying on to justify their irrational rise. Or perhaps they're not even relying on any justification - it's just floods of money looking for any old home, and generating and riding a self-reinforcing wave. These situations don't last long - it's either brutal correction or aggressive inflation.