Sic Semper Tyrannis

As part of his recent WSJ editorial, "Five Macroeconomic Myths", economist Ed Prescott "busts" the myth that the US public debt is "high", and in specific, burdensome to future generations.

In fact, Prescott argues that public debt is probably not high enough! Based on the models he believes are true, Prescott says the public debt should be around twice the nation's GDP (gross economic output) of about $13 trillion to optimize the welfare of current and future generations.

Herein I argue that, even accepting this premise, we should be deeply concerned about the public debt.

Follow up:

For the sake of argument, let's put aside the specific assumptions of Prescott's model (these aren't in the WSJ article, but have come to the surface in subsequent discussions in the blogosphere); they include an assumption of 4% real return on investment, 2% productivity growth, and 1% population growth -- all highly questionable. But let's assume Prescott is right about these things.

So we'll assume that Prescott's claim that a public debt twice the GDP is in some sense "optimal". Then how do we measure up, given the current situation?

Prescott thinks we have plenty of room to grow, as public debt is about 30% of GDP, at $4 trillion dollars, by his count.

When I saw this claim, alarm bells went off: last I checked the Bureau of Public Debt website, the Federal public debt was listed as more than $8 trillion.

Why the discrepancy? Prescott makes a mistake (or a fib, depending on how charitable you are) in leaving out fully half of this debt because half of it is in the form of "intragovernmental holdings". This means that the $4 trillion of the debt Prescott does not count is owed by one part of government to another (such as the Social Security trust fund).

But just because this debt is owed by the government to itself doesn't make it any less significant: in fact, this debt is really all owed to the American people, whether or not the bondholders are inside the government. I guarantee you people would notice if the Social Security Trust Fund were suddenly set to zero because the government decided to cancel the corresponding intragovernment holdings. Obviously, there would be real impact to doing this, so the debt is therefore quite "real".

So Federal public debt is at least $8 trillion, bringing us to a debt burden more like 60% of the GDP. This is still within Prescott's limit of 200%.

But we're not done yet! So far we've only discussed Federal debt. Anyone paying attention to The State of California's recent bond-flinging adventures has noticed that states have public debt too. Further, local municipalities also have debt. In fact it is this sort of debt that makes up the municipal bond sector, which many of you may have investments in (usually attractive because of it's tax-free nature).

Like Federal debt, state and local debt has approximately doubled in the past four years, and is now at about $2 trillion nationwide.

This brings the total notational value of public debt, at all levels nationwide, to about $10 trillion. Now we're up to about 76% of GDP.

But we can dig yet deeper: one might reasonably question the use of the GDP metric itself.

Not too long ago, we used a similar metric called the GNP, or Gross National Product. The change from GNP to GDP was done gradually and quietly over many years. Most people simply assume the two numbers are interchangeable; at most, differing by some inconsequential amount.

But this isn't true at all -- or rather it is only true in times where the deficit is very low. And this highlights the key difference between the two metrics: the old GNP subtracts international borrowings from the calculation of gross economic productivity, whereas the GDP actually adds them in! The very operation of borrowing a trillion dollars from overseas actually increases the measured GDP by a trillion dollars. This would be akin to you borrowing money on your credit card, chalking that up as income, spending it, and calling it a day. But that money isn't really income: it must be paid back. In fact, it even costs money to receive that money -- it is debt, after all.

So fundamentally, utilizing the GDP to determine the debt burden is illegitimate, because you're adding in the debt twice, both in the numerator (to the total debt number) and denominator (to the GDP)!

If we were to use the GNP instead, which is a more appropriate metric for measuring the debt burden, then we would have to shave about a trillion dollars in borrowings per year off the GDP number. This leaves us with a $12 trillion GNP, making the debt burden about 83%.

This is starting to look pretty steep, but is still not 200%.

Here's where the kicker comes in. As many sources, including the USA Today, reported earlier this year, if you add in total planned government liabilities to the mix (rather than just using current debt and liabilities -- basically, applying the same GAAP rules corporations must use), then the total debt load at all levels of government is more like $65 trillion.

This would make the debt burden more than 540%... well over Prescott's factor of two!

Further, if I am right and the GNP/GDP has been methodically overestimated by about 40% in the last 25 years due to inflation under-estimates, then the debt burden is in reality more like 900%!

Now, that's a pretty huge multiple, and the numbers seem fantastic. But they are in fact based on fundamentals: a major impact would certainly be felt in very real terms if the government was to cancel any debt, reneg on Social Security or Medicare promises, abandon the care of veterans, or fail to make good on any of it's other innumerable promises.

My conclusion is that, even if Prescott is right, we are already far beyond tapped-out and over-extended. And that's just looking at the public portion of the picture.

There are other problems with Prescott's argument. One is that we're already paying about $400 billion a year in interest on the Federal public debt, fully half of which does not go to intragovernmental holdings, and so is simply "wasted". This is money which does not go to providing actual public services, it goes entirely to holders of debt (a large portion of whom are foreign, these days). It is difficult to see how such waste, now making up the third-highest category of non-entitlement spending, could be considered "useful" for American citizens.

Finally there's a philosophical point: once money is lent to the government, it is difficult to control what it is spent on. While it is comforting to think of government debt as all going towards building roads and schools, in actuality it is thrown in one giant pool of funds which go towards many other things, including foreign wars and profligate "homeland security" spending. It is difficult to see how these things are any sort of prudent investment in the future, and Prescott's model is meaningless if large chunks of this public money are basically thrown away.

I conclude that we must already proceed with caution regarding the solvency of government. This is not the time to borrow even more and worry about it later -- not privately, and not publicly either. Sadly, we've probably already gone too far to avoid a painful "correction."

Addendum, Dec 18, 4:30pm: I've thought more about the fact that Prescott only addressed the optimality of the 2x debt burden level. This leaves a glaring open question: is it just as bad to be over the optimal level as under? That is, are we in as much peril if we have say a 300% debt burden as opposed to 100%? My guess would be no: a debt burden too high is likely to precipitate economic catastrophe, whereas a debt burden "too low" may simply have some vague ramifications about whether the economy is "growing as fast as it could" (that is, savings may be arguably "too high", as the establishment thinking goes).

Update, Dec 19, 11pm: As if to underscore this essay, the GAO released a report today reaffirming that the government is actually in the hole to the tune of $55 trillion. Close enough. Using my modifications to the national product, this gives us a debt burden in excess of 700%. By the "official" GDP number, the debt burden is at least 400%, still twice Walker's "safe" number. I also calculated today that if the government's weighted average cost of capital (the average interest rate of the debt) were to rise from the present approximately 4.5% to about 7% (as would conservatively be the case in high inflation or international abandonment of the dollar), the debt interest cost, even holding constant the present debt of about $9 trillion, would rise from today's $400 billion per year to around $650 billion per year. Just in interest -- and this is assuming none of this interest is ever added back into principal (negative amortization).